The Decline in Global Violence

Has the long-term threat of violence — war, terrorism, and homicide —  been decreasing or increasing worldwide? For some, the answer seems clear. Many in the strategic community concur with General Martin Dempsey, Chairman of the United States Joint Chiefs of Staff, who has said today’s world is “more dangerous than it has ever been.”  But there is little evidence to support them. During 2012, the number of conflicts being waged around the world dropped sharply, from 37 to 32. High – intensity conflicts have declined by more than half since the end of the Cold War, while terrorism, military coup and genocide numbers are also down. And this is not a recent phenomenon. According to Harvard University’s Steven Pinker, violence of all kinds has been declining for thousands of years; he has argued “we may be living in the most peaceful era in our species’ existence.” In the 2013 Human Security Report
The Decline in Global Violence: Evidence Explanation and Contestation — excerpted here — global security specialist Andrew Mack analyses the evidence. The report will be presented on March 3 to the United Nations and World Bank.





Published March 1, 2014

During the past decade, an increasing number of studies have made the case that levels of violence around the world have declined. Few have made much impact outside the research community — Steven Pinker’s The Better Angels of Our Nature: Why Violence Has Declined is a major exception.

Published in 2011, Better Angels’ central argument — one made over some 700 densely argued pages of text, supported by 70 pages of footnotes — is that there has been an extraordinary but little-recognized, long-term worldwide reduction in all forms of violence — one that stretches back at least to 10,000 BCE.

Better Angels has received high praise for its extraordinary scope, its originality, and the breadth and depth of its scholarship. It is engagingly written, powerfully argued, and its claims are supported by a mass of statistical evidence.

It has also generated considerable skepticism and in some cases outright hostility.

The Long-Term Decline in Violence

The decline in the violence that human beings perpetrate against each other has taken place in different periods in different parts of the world and there have been many reversals.

But the overall trend, Pinker argues, has clearly been downward — less warfare, fewer murders, dramatic reductions in torture and other cruel and inhumane practices, and the virtual eradication of slavery.

Better Angels is by far the most ambitious of the studies of trends in global violence that have appeared in the new millennium — including the Human Security Reports published in 2005 and 2011. Most of these studies have reported on reductions in the level of political violence — notably wars and terrorism — and have focused on the post–World War II world. The scope of Pinker’s study is much broader. Its historical sweep traverses some 12-plus millennia. It examines long-term declines in homicides as well as warfare, and a wide variety of forms of violence that are not necessarily lethal — slavery, rape and torture, and even cruelty to animals.

In explaining these remarkable changes, Pinker identifies five key trends.

First is the “Pacification Process” — the uneven transitions over thousands of years from anarchic hunter-gatherer, horticultural, and other early human societies to the first agricultural civilizations and then nation-states. These transitions have been associated with dramatic decreases in death rates from both war and homicides.

Second, from the late Middle Ages to the twentieth century came the “Civilizing Process” that accompanied the growth and consolidation of the nation-state system in Europe. During this period, Europe became more urban, more cosmopolitan, commercial, and secular. Often highly repressive, the Civilizing Process was associated with declines in homicide rates that ranged from tenfold to more than fifty-fold.

Third, the “Humanitarian Revolution” that started in the seventeenth and eighteenth centuries was associated with the decline and eventual abolition of slavery, with the slow elimination of judicial and other forms of torture and a long-term reduction in all manner of other cruel and inhumane practices.

Fourth, the“Long Peace”that followed the end of World War II saw the disappearance of great-power wars and the dramatic reduction in the number and deadliness of other international conflicts. This change came about in part because industrial-strength warfare had become so destructive to all parties that it no longer served any rational purpose.

The popular revulsion generated by the mass slaughters of World War II had also strengthened the emergent norm that proscribed the resort to war except in self-defence or with the imprimatur of the UN (United Nations) Security Council. By the early 1970s, wars of liberation from colonial rule were mostly over and the idea of new colonial conquests had become simply unthinkable.

Finally came what Pinker calls the“New Peace”of the post–Cold War period. From the early 1990s, the number of conflicts within states declined substantially after increasing for some four decades.

The end of the Cold War not only removed a significant source of conflict from the international system, it also led to the emergence of a new form of global security governance.

Starting in the early 1990s, the much-criticized UN spearheaded a massive upsurge of international activism directed towards preventing wars, stopping those that could not be prevented, and preventing those that had stopped from reigniting. Its key stakeholders have been international agencies, donor governments — and those of war-affected states — plus huge numbers of NGOs (non-governmental organizations).

In its current stage of development, this continually expanding system of global security governance remains inchoate, disputatious, inefficient, and prone to tragic mistakes. But as previous Human Security Reports have argued, the evidence suggests that it has also been remarkably effective in driving down the number and deadliness of armed conflicts.

“Better Angels”and“Inner Demons”

 Steven Pinker is an experimental psychologist and cognitive scientist, so it is not surprising that he devotes an entire section of Better Angels to the psychological mechanisms that drive violence. Human beings, he argues, are neither innately good nor evil, but circumstance can orient them either towards confrontation and violence or towards cooperation and peace. The drivers of violence — predation, dominance, revenge, sadism, and ideology — are the “inner demons”of human nature. The “better angels of our nature” — the faculties that steer individuals away from violence — include empathy, self-control, moral sense, and reason. (The references to the “inner demons” and “better angels” of human nature come from United States President Abraham Lincoln’s first inaugural address, delivered in 1861.)

Better Angels can be understood in part as an analysis of how long-term changes in culture and material circumstance have, over time, permitted the better angels of human nature to prevail over its inner demons.

Contested Claims


Reenactment at the Viking Festival in Newfoundland, celebrating the 1,000 year-old Norse settlement at L’Anse aux Meadows. Photo by Greg Locke Copyright © 2013

Most reviews of Better Angels have been highly positive, though some otherwise sympathetic reviewers have challenged particular claims, especially those regarding the deadliness of major episodes of violence. But some critiques have been consistently negative and a few have been deeply hostile.

The focus of the most sustained criticism has been Pinker’s central claim — that there has been a millennia-long decline in all forms of violence.

Claims that the number of interstate wars has decreased dramatically since the 1950s, and that civil war numbers have declined since the end of the Cold War, are now uncontroversial within the mainstream conflict research community, though they still occasion surprise and sometimes skepticism among non-specialists. The data on violence going back 10,000-plus years are far less robust and it is here that Pinker’s analysis is most audacious — daring to tread where few scholars have gone before and generating some intense criticism in so doing.

Better Angels is criticized for underestimating the violence of today and for overestimating that of the distant past.

A smaller number of critics accept that there have been significant reductions in violence over the centuries but challenge the complex multi-level arguments Pinker advances to explain them.

Those who reject the evidence that is marshalled in Better Angels fall into two broad camps. Against Pinker, the majority affirm the conventional wisdom that World War II was, in fact, the deadliest-ever conflict and that the twentieth century was the most violent in history. A smaller number of critics —  mostly anthropologists — argue that the hunter-gatherer and other societies that preceded the formation of states were far less violent than Pinker claims.

Better Angels, in other words, is under attack, both for underestimating the violence of the recent past and for overestimating that of the distant past. Thus, to sustain his thesis that there has been a millennia-long decline in violence, Pinker has to do two things. First, he has to argue that World War II was not the bloodiest conflict in world history. Second, he has to show that the anarchic hunter-gatherer and other non-state groups that made up the earliest human societies had far higher rates of lethal violence than the state-based societies that succeeded them.

Was World War II the Deadliest War in History?

One reason that the core thesis of Better Angels has invited so much surprise — and skepticism — is that, in terms of the sheer numbers of people killed, there is little doubt that World War II’s death toll was greater than any other war in the entire span of human history. This uncontested fact raises an obvious question: if the deadliest-ever war took place in living memory, how is it possible to claim, as Pinker does, that we are likely living in the least violent era in human history?

Pinker does not dispute the fact that World War II almost certainly killed more people than any other war in history. But he argues that the most appropriate metric for estimating the deadliness of wars is not the absolute number of fatalities but the number of war deaths relative to the size of the population. From this perspective, a conflict that kills 10,000 people in a society with a population of 100,000 is 10 times deadlier than one that kills 10,000 people in a society of a million people even though the numbers killed are identical.

While World War II certainly killed far more people than did earlier episodes of mass killing, the global population was far larger in the twentieth century than in earlier centuries, making World War II’s bloodletting relatively much less deadly than the absolute numbers suggest. Indeed, Pinker maintains that relative to the world’s population, World War II is only the ninth-deadliest episode of sustained violence in human history.

But this is not the metric most commonly used for determining the deadliness of periods of violence. The standard measure is deaths per 100,000 of the population per year. Because the huge number of World War II deaths occurred within a very short period, the annual rate of killing was far greater than in earlier episodes of mass violence, most of which occurred over far longer time periods. Indeed, using the standard metric of violent deaths per 100,000 of the population per year, World War II becomes the deadliest war in more than 1,000 years.

However, if we take a longer time horizon and turn to yet another metric, the picture changes again. The quantitative data that Pinker draws on for deaths caused by violent conflict in pre-historic and other early non-state societies indicate that, on average, warfare accounted for about 15 percent of fatalities from all causes. This is an astonishingly high rate — dramatically greater than the percentage of deaths caused by warfare in modern Europe — even in the two deadliest centuries of the most recent half millennium. Thus, wars in seventeenth-century Europe were responsible for some two percent of deaths from all causes; in the twentieth century, the figure was three percent — one-fifth the average rate of the early hunter-gatherer societies. Focusing on the longer period is important because, as Pinker makes it clear, his thesis is about a global decline in violence that covers the period from pre-history to the present day — not simply Europe since the Middle Ages.

How Violent Were Early Human Societies?

To sustain his declinist argument, Pinker has to argue not only that the twentieth century was relatively much less deadly than the conventional wisdom suggests but also that the human societies that existed millennia ago were far more violent than widely understood. This latter claim has come under fierce attack from anthropologists who reject the quantitative data — much of it archaeological — on which Pinker draws. The critics argue that the data from these periods are sometimes wrong, and that they are too few and fragmentary to have confidence that they are representative of levels of violence in all non-state societies.

In a new study released by Oxford University Press in 2013, editor and leading Pinker critic Douglas Fry maintains that there is virtually no evidence that the earliest nomadic societies, those existing prior to 10,000 BCE, were warlike. Pinker’s declinist thesis fails, Fry argues, because it ignores the fact that the earliest human societies were extraordinarily peaceful.

However, the data from this very early period are even scarcer and more fragmented than those of more recent millennia. And, of course, the absence of evidence for high-fatality warfare in this period is not the same as evidence for the absence of such violence.

Moreover, even if Fry’s argument about the peacefulness of the very earliest human societies is accepted, it is far from clear how this undermines Pinker’s thesis. The focus of Better Angels is predominantly on the period after 10,000 BCE, and Fry agrees both with Pinker’s contention that violence in this latter period decreased and that, over time, the spread and consolidation of state power had a pacifying effect on inter-group violence.

Lethal Violence Is about More Than War Deaths

Mostar wall

A wall in Mostar, Bosnia and Herzegovina. Slowly, graffiti camouflages the destruction of the Balkans war. Photo © Deborah Jones 2012

The controversies over which periods of warfare in human history are the deadliest are unlikely to be resolved any time soon. However, even if we had long-term war-death data that were comprehensive and reliable, this in itself would not be sufficient to either prove or disprove Pinker’s thesis. Better Angels is not just about millennia-long trends in warfare but all forms of violence, not least homicides. This is pertinent because as the Geneva Declaration on Armed Violence has demonstrated, warfare is responsible for less than one in 10 violent deaths in today’s world — the large majority result from homicides.

The available data from Western Europe reveal a dramatic decline in homicide rates over the past 700-odd years. Pinker draws on a range of quantitative studies indicating that the average homicide rate across the region fell from well over 50 per 100,000 per year in the fourteenth century to little over one per 100,000 in the twentieth century, a fiftyfold decline. Over the same period, the limited data cited in Better Angels indicate that deaths per 100,000 of the population per year from great- power wars increased as states consolidated, grew larger, and the killing power of their armies expanded.

We cannot, however, assume that war death rates overall necessarily increased in this period because it is likely that death tolls from civil wars decreased as state control expanded and consolidated, while death tolls from great-power wars became more deadly. It is also possible that any net increase in overall war death rates was more than offset by the steep decrease in average homicide rates over the same period.

While focusing on changes in the incidence of deadly violence over 700 years of European history is instructive, it is again important to remember that the total period that Pinker examines ranges back more than 12,000 years. Here the mass of evidence marshalled in Better Angels indicates that as human societies transitioned during this period from hunter-gatherer and hunter-horticulturist societies to those governed by states, death rates from violence declined dramatically. As Pinker points out: “Modern Western countries, even in their most war-torn centuries, suffered no more than around a quarter of the average death rate of non-state societies, and less than a tenth of that for the most violent one.

A very similar trend is evident for homicide rates in the transition from non-state to state-based societies.

Finally, we note that the violence described in Better Angels includes a wide range of violent practices that do not necessarily kill their victims and thus are not counted in the fatality datasets. These include slavery, torture, cruel and inhumane punishments, and the physical abuse of children and the mentally ill. Most of these practices have been eliminated, proscribed, or greatly reduced in recent centuries. These changes have not been challenged by any of the critics of the declinist thesis.

Is Organized Criminal Violence Becoming a Greater Threat Than War?

 The large majority of countries in the world are not plagued by wars, while all suffer lethal criminal violence, so it is not surprising that far more people worldwide die as a result of homicides than warfare.

Most of the focus in the declinist literature has been on war, yet relatively little attention has been paid to homicides. But, like Steven Pinker, the researchers associated with the Geneva Declaration on Armed Violence have made a strong case for looking at trends in all forms of lethal violence. The period that they focus on does not, however, encompass millennia but simply the few recent decades for which the UN has global homicide data.

Mexico’s homicide rate, while high, is actually lower than those of other regional states that are also deeply affected by drug- related organized crime.

The analyses of the Geneva Declaration researchers have drawn attention to what Steven Zyck and Robert Muggah have described as the “growing scale and significance of chronic organised criminal violence, often sustained by trans-national crime networks.”4 Nowhere has this trend been more evident than in Mexico where, in 2011, the death toll from drug-related organized criminal violence was higher than the battle-death toll of the war in Afghanistan or Sudan or Iraq.

When we look at homicide rates, rather than absolute numbers, we find that Mexico’s rate per 100,000 of the population per year was considerably lower than those in four other Central American states that have also been deeply affected by drug-related organized crime. And homicide rates associated with organized criminal violence were not just high —  in the first decade of the twenty-first century they grew substantially, not just in Mexico but also in Guatemala, El Salvador, Honduras, and Belize.

So, while the civil wars that plagued much of Central America were mostly over by the early 1990s, a deadly new form of organized killing appeared to have replaced the lethal violence of warfare.

Organized and drug-related criminal violence is not, of course, restricted to Central America. It also afflicts Afghanistan, Southeast Asia’s “Golden Triangle,” and parts of West Africa, the Caribbean, South America, and southern Europe. This raises an interesting question: have there been dramatic increases in organized criminal violence in these regions as well? And if so, might any such increases be greater than the uneven decline in violent war deaths?

The United Nations Office on Drugs and Crime (UNODC) cannot provide a definitive answer to this question because the homicide data it receives from national governments around the world rarely distinguish between homicides perpetrated by criminal organizations and the far more numerous “individual” homicides that take place, most of them within families or between acquaintances.

The Uppsala Conflict Data Program’s (UCDP’s) armed conflict dataset on which Pinker, the Human Security Report, other researchers, and many international agencies rely can, in principle, provide an answer to this question. Since 1989 UCDP has been seeking to track all forms of organized violence — criminal as well as political. (It does not track individual homicides.) However, in practice, as we point out in Chapter 2, UCDP’s stringent data coding rules mean that a large percentage of the deaths that result from organized criminal violence in Mexico and elsewhere cannot be coded and therefore do not get recorded.

Our review of the very limited data on organized criminal violence around the world concludes that the increasing and extraordinarily high levels of lethal violence perpetrated by organized gangs in Mexico and Central America in the first decade of the twenty-first century are not representative of the rest of the world.

 Moreover, even in this region there are signs of change. The number of homicides attributable to organized crime in Mexico declined by some 28 percent from 2011 to 2012 according to the Mexican government.

In Guatemala the murder rate declined substantially between 2009 and 2011. In El Salvador a 2012 agreement signed between rival organized crime gangs led to a sharp decline in the murder rate. In the first half of 2013, the number of homicides dropped by one-third compared with the first half of 2012.

There is, of course, no guarantee that these encouraging declines will continue, but at the very least they serve to remind us that there is nothing inevitable about high and rising homicide rates in countries afflicted by drug-related organized crime. The 50-percent-plus drop in Colombia’s homicide rate that the UN recorded for the period between 2002 and 2010 is a further reminder of how quickly crime rates associated with drug trafficking can decline in the Americas. By the end of 2012, the murder rate in Colombia — until recently the major source of the world’s cocaine supply — was at its lowest in 27 years.

Are Global Homicide Rates Increasing Overall?

There is no doubt that war deaths from injuries have decreased dramatically since the end of World War II, but we have little idea about global trends in homicides.


Young soldiers in Uganda. Photo by Greg Locke Copyright © 2013

For most of the World War II period there are no reliable homicide data for developing countries. Even today the data for some countries — most of them in sub-Saharan Africa —  are either non-existent or highly unreliable.

It is possible, then, that homicides could have increased worldwide since World War II, while war deaths declined. The 2011 Global Burden of Armed Violence (GBAV) report, for example, found that between 2004 and 2009, the worldwide non-conflict homicide rate increased by some 5 percent. But no conclusions — particularly about long-term trends —  should be drawn from the world homicide data compiled by UNODC.

This is because, until recently, many poor-country governments did not report homicide data to UNODC. Times are changing, however, and every year more national statistics offices in the developing world are collecting and reporting homicides to the UN. This matters because, on average, low-income countries tend to have higher homicide rates than medium- and high-income countries, so adding more poor-country data to the UNODC’s global homicide database will have the effect of increasing the average global homicide rate. In other words, the increase that the GBAV researchers recorded between 2004 and 2009 may well be a function of more reporting of homicides rather than more actual homicides.

But there is a more compelling reason for being skeptical that homicide rates have been increasing worldwide, and not just between 2004 and 2009. As *** Chapter 2 points out, on average, medium- and high-income countries tend to have substantially lower homicide rates than low- and low-medium-income countries. We would expect, therefore, that as income levels rise on average in the developing world — as they have done substantially since the end of World War II — homicide rates would tend to fall.

Any such declines will likely be driven by the growth and consolidation of the power of national governments — i.e., essentially the same factors that Pinker argues drove the “pacification”and “civilizing”processes in Europe and elsewhere. Higher incomes do not in themselves cause violence to decline, of course, but rather they are associated with increased state capacity. In practice this means that as national incomes increase, states have more resources to deter, stop, and otherwise prevent violent crime.

Taking Stock and Looking Ahead

Clearly, the key findings presented in Better Angels are contested. This is not surprising since the data that could resolve many of the disputed factual claims, particularly from past centuries in what we now call the developing world, are either non-existent or too sparse and fragmented to reliably indicate global trends. And even when there is no doubt about the global decline in a particular form of violence over a particular period — as is the case with the reduction in the number and deadliness of interstate wars after World War II, for example — the causes of that decline can be — and are — disputed.

The most arresting findings in Better Angels are those that point to the extraordinary millennia-long declines in homicide and war death rates that have been associated with the transitions from hunter-gatherer and horticulturalist societies to those governed by states. There is, as noted previously, some controversy about the mortality data from violence from the very earliest hunter-gatherer societies. But even Douglas Fry, the most prominent critic of Better Angels’ analysis of the pre-historic era, does not dispute the data that indicate that lethal violence declined after 10,000 BCE as the anarchy of nomadic non-state societies began to be replaced by the spread of early civilizations and state-based systems of social control. The remarkable trend data that Pinker has collated on homicide rates in Europe from medieval times to the present lend further support to the declinist thesis. These data have not been challenged.

The case for pessimism about the global security future is well rehearsed and has support within the research community.

The most encouraging data from the modern era come from the post–World War II years. This period includes the dramatic decline in the number and deadliness of international wars since the end of World War II and the reversal of the decades-long increase in civil war numbers that followed the end of the Cold War in the early 1990s.

What are the chances that these positive changes will be sustained? No one really knows. There are too many future unknowns to make predictions with any degree of confidence. And Pinker makes it very clear that his thesis seeks to explain the decline of violence in the past, not to predict the future. Moreover, the case for pessimism about the global security future is well rehearsed and has considerable support within the research community. Major sources of concern include the possibility of outbreaks of nuclear terrorism, a massive transnational upsurge of lethal Islamist radicalism, or wars triggered by mass droughts and population movements driven by climate change.

Pinker notes reasons for concern about each of these potential future threats but also skepticism about the more extreme claims of the conflict pessimists. Other possible drivers of global violence include the political crises that could follow the collapse of the international financial system and destabilizing shifts in the global balance of economic and military power — the latter being a major concern of realist scholars worried about the economic and military rise of China.

But focusing exclusively on factors and processes that may increase the risks of large- scale violence around the world, while ignoring those that decrease it, also almost certainly leads to unduly pessimistic conclusions.

In the current era, factors and processes that reduce the risks of violence not only include the enduring impact of the long-term trends identified in Better Angels but also the disappearance of two major drivers of warfare in the post–World War II period — colonialism and the Cold War. Other post–World War II changes that have reduced the risks of war include the entrenchment of the global norm against interstate warfare except in self- defence or with the authority of the UN Security Council; the intensification of economic and financial interdependence that increases the costs and decreases the benefits of cross- border warfare; the spread of stable democracies; and the caution-inducing impact of nuclear weapons on relations between the major powers.

With respect to civil wars, the emergent and still-growing system of global security governance has clearly

helped reduce the number of intrastate conflicts since the end of the Cold War. And, at what might be called the “structural”level, we have witnessed steady increases in national incomes across the developing world. This is important because one of the strongest findings from econometric research on the causes of war is that the risk of civil wars declines as national incomes — and hence governance and other capacities — increase. A remarkable recent statistical study by the Peace Research Institute, Oslo (PRIO), that found that if current trends in key structural variables are sustained, the proportion of the world’s countries afflicted by civil wars will halve by 2050.

Copyright © 2014 by Human Security Research Group

Published by F&O with permission from Andrew Mack and the Human Security Research Group.

Andrew Mack is the Director of the Human Security Report Project (HSRP) at Simon Fraser University in Vancouver, and a faculty member of the university’s new School for International Studies. Professor Mack has also held research and teaching positions at Harvard, the University of British Columbia, Australian National University, among many other academic institutions. From 1998 to 2001 he was Director of the Strategic Planning Office in the Executive Office of UN Secretary-General Kofi Annan. His pre-academic career included six years in the Royal Air Force (as an engineer and pilot); two and a half years in Antarctica as a meteorologist and deputy base commander; a year as a diamond prospector in Sierra Leone; and two years with the BBC’s World Service writing and broadcasting news commentaries and producing the current affairs program, The World Today. Professor Mack has written and edited some 16 monographs and books and his 60-plus scholarly articles have appeared in a wide range of journals. He is widely published in mainstream print media.

Further reading:
Human Security Report Project


Support professional journalism at Facts and Opinions: Use the subscribe form sign up for email notices from Frontlines, where we post small stories and announce new work.  Click here to purchase a site day pass or subscription.