Is humanity becoming less violent?

Over centuries, has everyday violence actually fallen?

Homicide has fallen for seven centuries.

In a medieval town of forty thousand, ten people might be murdered in an ordinary year, and in the worst decades nearer fourteen. A town that size today could go three or four years between killings. A medieval European was something like thirty-five times more likely than you to die at another’s hand — a gap so wide it is less a number than a different climate of life, one in which sudden violence was part of the weather.

What closed it was not a change in the human heart but in who was allowed to answer an insult. Over centuries, kings and courts took the settling of scores out of private hands; the feud and the duel gave way to the magistrate, and killing a rival stopped being a point of honour and became a crime you could hang for. That is how Manuel Eisner, the criminologist whose court records run beneath this chart, reads the long descent: fewer men reaching for a blade, more of them reaching for the law. The small rise since its low around 1950 is the reminder underneath it all — this was hard-won, and it is reversible, never a law of nature.

Across the centuries, has war grown more or less deadly?

War’s deadliest century is the one we just left.

Measured against the number of people alive, the twentieth century was the most violent in six hundred years of record. The two world wars drove the global death rate from armed conflict to a peak in 1941, over 300 deaths a year for every 100,000 people, a level with no rival going back to 1400. Even the wars of the 2020s have pushed today’s rate only to around two.

Then it broke the other way. After 1945 the rate collapsed and stayed low for the rest of the century, the long stretch historians call the Long Peace: no war between the great powers, and nothing close to the world wars’ toll. The line is two stories at once. War became, for a while, more industrial and more lethal than ever, and then more avoided than ever. Which of those is the real trend is what the last few years are testing. The line stops in 2024, but the story has not: Ukraine pushed the rate sharply upward in 2022, and the wars still burning in 2025, in Sudan, Gaza, and the Sahel, have held it well above the lull of the 2000s. The charts that follow carry that recent toll in full.

Are the world’s wars killing fewer people?

War deaths fell far below their Cold-War peaks — then turned back up.

In 1950, about half a million people died in battle, most of them in the Korean War. The toll then fell for half a century, down to roughly twelve thousand by 2005, a forty-fold drop that looked like proof the world had left mass warfare behind.

That fall had a cause worth naming. After 1945 the great powers stopped fighting each other head-on, held back partly by the fear of nuclear weapons, and the wars that remained were mostly smaller civil and proxy conflicts. What the right of the chart shows is that arrangement straining: Syria, then Ethiopia’s Tigray war, then Russia’s invasion of Ukraine and the war in Gaza have pushed the annual toll back above a hundred thousand. One thing the line hides is that it counts only deaths in combat, not the larger numbers who die from the hunger and disease that war drags behind it.

How much did the last five years actually undo?

Three wars reversed two decades of progress.

In 2010, about 21,000 people died in the world’s wars. Then came Syria, Yemen, Ethiopia, Ukraine, and Gaza. By 2022 the annual toll had hit 276,893, and even after easing it sat above 128,000 in 2024. That is still well short of the 550,000 killed in 1950, but the line on this zoomed-in view climbs more than tenfold in a decade. The long decline historians celebrate is real; these last years show how fast it can be undone.

If deaths fell, why does the world feel more violent?

There are more active conflicts now than at any point in the record.

In 2010 the world had 31 state-based conflicts running, the fewest in the modern record. By 2024 it had 61, the most in the thirty-six years anyone has counted this way; the number nearly doubled even through stretches when the death toll was falling.

Two things drive the multiplication. Insurgencies spread: the jihadist war that began in Mali in 2012 now burns across its neighbours, and every border it crosses adds another conflict to the ledger. And outside powers keep joining other people’s civil wars, with money, drones, and mercenaries, which is precisely what makes wars longer and harder to settle. So a falling death toll and a rising conflict count are not a contradiction. The world is not always fighting bigger wars. It is fighting more of them, in more places, with more hands in each.

If the death rate is falling, why are the human totals rising?

More people are refugees now than at any time on record.

In 1960, the world counted 150,000 refugees. Today the figure has crossed 30 million, a two-hundred-fold rise, and it has roughly tripled just since 2010. The surge is not mainly that the world grew more violent per person; it is that more states have shattered and more people have somewhere worse to flee. The death rate from violence has fallen. The number of people forced across a border to escape it has never been higher.

Where are the world’s wars actually killing people?

War never spreads evenly. It concentrates, and the place keeps moving.

Follow the deadliest place to fight a war and it keeps moving. In the mid-2010s it was the Middle East, where Syria and the war against ISIS killed more than 80,000 people in 2014 alone. Then the centre of gravity slid to Africa: Ethiopia’s Tigray war helped make 2022 the deadliest year on this chart, with around 175,000 killed. By 2024 it had moved again, to Europe, where the trenches of Ukraine were taking more than 100,000 lives a year.

This is what a flat global line hides. The world’s total can hold almost steady while underneath it the killing migrates from one region to the next, each taking its turn to carry most of the dead. For the people living through it there is no global average — only the war on their own doorstep, which in that moment is the entire world. The map of war is never blank; it just keeps redrawing where the darkest patch falls.

What are the world’s wars actually killing people right now?

A cluster of wars broke out in the 2020s, and most are still burning.

Name the deadliest wars of the decade and you have named most of the recent rise. Ethiopia’s Tigray war was, for a moment in 2022, the bloodiest on earth: more than 160,000 people killed in a single year before a ceasefire ended it. Russia’s invasion of Ukraine has been the most relentless, taking close to 100,000 lives a year since 2022 with no end in sight. Since 2023, Sudan’s civil war and Israel’s war in Gaza have each killed tens of thousands more, while a jihadist insurgency keeps spreading across the Sahel.

What makes this stretch dangerous is not any single war but the pile-up. For two decades after the Cold War the world rarely had more than one conflict killing at this scale; now several burn at once, on three continents, and the chart names only the largest. Beyond it, eastern Congo is sliding back toward full-scale war, Myanmar’s junta is losing a civil war almost no one watches, and al-Shabaab still bleeds Somalia. A few are winding down: Syria’s long war collapsed when Assad fell in late 2024, and the Israel-Iran strikes of 2025 flared and then stopped short of the regional war everyone feared. The through-line is a world with too many fires lit at once, and not enough being put out.

What about the wars no state is fighting?

Between war and murder sits a violence the categories barely name.

When two cartels fight over a smuggling route, no state is at war and no murder is ordinary. Researchers call it non-state conflict: organised armed groups killing each other, with the government on neither side. In 1989 this kind of fighting took about 4,200 lives worldwide. In 2017, its worst recorded year, it took just over 31,000, much of it in Mexico’s cartel wars and in militia fighting across Africa.

It is the violence our categories are built to miss. Too organised to sit among ordinary murders, too stateless to count as war, it fell between the lines of both for decades, which is why this chart starts only in 1989. The toll has eased since the peak, to around 14,500 last year, but it never goes quiet. Whether the world is at peace depends a great deal on whether you count the wars nobody declares.

Is “the homicide rate” even a single number?

Where you live changes your odds more than anything else about you.

Wars cluster in a handful of places, and so does ordinary murder. There is no such thing as the homicide rate. In Latin America and the Caribbean, around twenty people per 100,000 are murdered every year; in East Asia and the Pacific, fewer than one. That is a twenty-fold gap between whole regions, and country to country it stretches past a hundredfold. Your odds of being killed are set more by where you were born than by almost anything else about you.

These are recent numbers, and they cut against the comfortable global average of about five. Latin America’s toll is driven by organised crime and overwhelmed courts, not war; Sub-Saharan Africa sits high too, while Europe and East Asia have pushed murder to the edge of statistical noise. The single world figure everyone quotes is a fiction no one actually lives in. It averages Caracas into Tokyo, and so describes neither.

Who does the remaining violence actually kill?

The world’s murder victims are overwhelmingly men.

The average homicide rate hides who is dying. Worldwide, men are killed at about four times the rate of women: roughly eight per 100,000 against two. Most of those men are young, and most are killed by other men, in fights, feuds, and the business of organised crime. The single global figure averages a relatively safe group and a far more dangerous one into a number that fits neither.

This is not the whole of gendered violence, only its lethal, public part. Women are far more likely to be killed at home, by a partner or a relative, than by a stranger in the street, a kind of killing the headline rate barely distinguishes. So the chart says something true and narrow: in the open, among strangers, it is mostly men who kill and men who die. The violence aimed at women tends to happen behind a door the data struggles to see.

Has humanity stepped back from the ultimate violence?

The world’s nuclear stockpile is a fifth of its Cold-War peak.

In 1945 the world held six nuclear warheads, all American. By 1986 the two superpowers had built 70,374 between them, enough to end civilization several times over. Since then the stockpile has fallen to roughly 12,000, about a fifth of the peak.

This is the largest deliberate disarmament in history, and it had names and dates: the end of the Cold War, and a run of arms-control treaties from the 1987 INF agreement to START, under which the United States and Russia cut the arsenals that make up most of the world’s warheads. The reduction was political, not technological; the weapons never became less terrible. And it has stalled. Twelve thousand warheads is still enough to kill hundreds of millions, and what remains is being modernised rather than retired.

Is the world disarming, or just rearming differently?

Even as warheads fell, military spending hit an all-time high.

In 1998, the world spent about 1.2 trillion dollars on armies, navies, and air forces. That was the low point after the Cold War. By 2025, the bill had climbed past 2.7 trillion, an all-time high. Warhead counts fell, treaties were signed, but the money never stopped flowing. The planet is not disarming; it is spending more than ever, just on different things — drones, cyber, special forces, the quiet machinery of modern conflict.

What about the violence aimed at civilians on purpose?

Sometimes the killing has no battlefield.

In a hundred days in 1994, close to 780,000 people were killed in Rwanda — not in battle, but hunted down, mostly with machetes, for belonging to the wrong group. That single year towers over everything else on this chart. One-sided violence is the deliberate killing of people who cannot fight back, and at its extreme it is the arithmetic of genocide.

This kind of killing has no front line and no symmetry: a state or a militia on one side, unarmed civilians on the other. In most years the global toll runs in the low tens of thousands, far beneath the Rwandan peak, and yet it never settles at zero. It rises whenever a government turns on its own people, or the losing side of a war is made to pay, and it has climbed sharply again these last two years. The capacity for massacre did not die in the twentieth century. It only went quiet.

Does the violence we fear most match what actually kills us?

Terrorism kills a rounding error of the violent dead — and rules the fear.

In 1970, 174 people died in terrorist attacks worldwide. In 2014, the deadliest year on record, the toll was 44,581. That sounds like a lot until you set it beside the rest of this article: the wars of that same year killed more than twice as many people, and everyday homicide quietly takes more lives than either. Terrorism dominates headlines, security budgets, and the fear of violence, yet for any one person the risk is vanishingly small. The gap between what we fear and what actually kills us is its own kind of damage, the one that warps how we spend money, attention, and freedom.