Read our COVID-19 research and news.

A new study suggests that we might not be out of the woods when it comes to large-scale war. Here, a soldier walks through London’s Leicester Square after it was bombed by the Germans during World War II.


Are we in the middle of a long peace—or on the brink of a major war?

Since the close of World War II, humanity has seen few large-scale wars—and battlefield deaths—compared with the past 2 centuries. War scholars refer to our current era as the “long peace.” But are we really getting along any better? A new study argues it will take another 100 years to see whether we are—or whether our relative peace is just the middle of a statistical blip.

Anthropologists and political scientists have argued that several 20th century developments have lessened the risk of large-scale interstate conflicts, including the spread of democracy, increased economic interdependence, and the threat of mutual nuclear annihilation. From a purely numerical perspective, those arguments carry weight. Only a handful of recent conflicts have boiled over into all-out war, and few have seen the kinds of extraordinary body counts—in the hundreds of thousands—common to wars in the 19th and 20th centuries.

To find out whether we are really in the midst of a period of true harmony, computer scientist Aaron Clauset at the University of Colorado in Boulder plotted data from The Correlates of War Project’s data set, widely used in political science. The database includes the year of onset and number of battle deaths for 95 interstate conflicts between 1823 and 2003. The Crimean War, for example, kicked off in 1853 and eventually killed about 264,000 soldiers in battle. World War II killed more than 16.6 million. And the Yom Kippur War killed some 14,400 soldiers in 1973. Data on civil wars and nonstate actors, such as terrorist groups, were not included in this analysis.

Clauset crunched the data as if they represented any other statistical relationship, looking for trends and calculating the normal range of fluctuation in both battle deaths and years between conflicts. Numbers of battle dead for a given war ranged from 1000—the minimum number in the data set—to the millions killed in World War II. For the purposes of the study, Clauset defined a large war as one whose battle death toll falls within the upper quartile of total battle deaths over the past 2 centuries. Basically, that means anything with more than 26,625 deaths.

Then he developed a series of computer models to replay the period from 1823 to 2003, using only statistical likelihood from the data to determine the frequency and severity of imaginary interstate wars. For one thing, Clauset found the unimaginable carnage of World War II was not, in fact, a statistical anomaly; its death toll falls well within the expected range for war deaths.

But he also found that, statistically speaking, going several decades without a large war simply isn’t a rare event—and that peace can quickly be upended by another large conflict. From a statistical perspective, there’s nothing special about the current “long peace,” Clauset reports today in Science Advances. In order for our present peaceful era to become meaningfully aberrant—that is, for it to represent a real change in our ability to get along—it would have to last for another 100 to 140 years.

What’s the statistical risk that we’re due for another large war sometime soon? “Not small,” Clauset says, though the number of unknowable variables about the future make it difficult to forecast with any confidence.

“The so-called ‘long peace’ trend might not be that long, and it might not be a trend,” says anthropologist Rahul Oka of the University of Notre Dame in Indiana, who was not involved in the study. “This is a great first step for this line of research.” Oka noted that future studies would benefit from running the numbers not only for battle deaths, but also the sizes of the armies involved.

There is some reason to be optimistic, however. Across all the models’ runs, a period of great violence (like the two world wars) followed immediately by a long span of relative peace appears to be genuinely rare, occurring in fewer than 1% of simulations. Maybe the peaceniks are onto something, after all.

*Correction, 22 February, 10:55 a.m.: An earlier version of this story misstated the number of battle deaths in the Yom Kippur War.