Two Projects Aim To Predict Which Countries Have The Greatest Risk Of Genocide : Goats and Soda Scholars at the U.S. Holocaust Memorial Museum say that mass killings follow predictable patterns. They're using a computer model to track where the next genocide is likely to occur.

Is Genocide Predictable? Researchers Say Absolutely

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript


Here's a question that still plagues humanity. How do you keep one group of people from slaughtering another? The U.S. Holocaust Memorial Museum is trying to find an answer. They are leading an effort called the Early Warning Project to try to identify where mass killings may be likely to occur. NPR's Jason Beaubien reports.

JASON BEAUBIEN, BYLINE: History, unfortunately, does repeat itself. Two thousand years ago, the Romans laid siege to Carthage, killing more than half of the city's residents and enslaving the rest. Hitler attempted to annihilate the Jews in Europe. In 1994, the Hutus turned on the Tutsis in Rwanda. The Khmer Rouge killed a quarter of Cambodia's population. After the breakup of Yugoslavia, the Serbs slaughtered thousands of Bosnians at Srebrenica in July of 1995. In 2017, when Buddhists attacked Rohingya Muslims in Myanmar, many people were shocked to hear that mass killings still occur in the 21st century. But there's growing evidence that these events follow familiar patterns.

JILL SAVITT: Genocides are not spontaneous.

BEAUBIEN: That's Jill Savitt. She's the acting director of the Simon-Skjodt Center for the Prevention of Genocide at the U.S. Holocaust Memorial Museum in Washington, D.C.

SAVITT: In the lead-up to these types of crimes, we do see consistent set of things happening.

BEAUBIEN: The Holocaust Museum, along with scholars from Dartmouth, has mapped the conditions that precede a genocide. They built a database of every mass killing since World War II. Then, they went back and looked at the conditions in those countries just prior to the attacks. And now they use that computer model to analyze which nations currently are at greatest risk.

SAVITT: This is not - we're not forecasting with precision. That's not the intention of the tool. What we're doing is trying to alert policymakers that here's a situation that is ripe for horrors to happen.

BEAUBIEN: In the three years prior to the attacks on the Rohingya, Myanmar ranked as the country most likely to have a mass killing for two of those years and ranked No. 3 in the other. The museum's computer model analyzes statistics that you might think have nothing to do with genocide - fluctuations in per capita gross domestic product, infant mortality rates, overall population size. They also plug in data about recent coup attempts, levels of authoritarianism, civil rights, political killings and ethnic polarization. Lawrence Woocher, who works on the Early Warning Project, says the form of government is one of the key data points. The most dangerous appears to be a government that's not a full dictatorship nor a full democracy.

LAWRENCE WOOCHER: The, I think, prevailing view about why mass atrocities occur is that they tend to be decisions by political elites when they feel under threat and in a condition of instability. And there's lots of analysis that suggests that these middle-regime types are less stable than full democracies or full autocracies.

BEAUBIEN: The project ranks 162 countries by their potential for a new mass killing to erupt in the coming year. The Democratic Republic of Congo is currently the most at risk, followed by Afghanistan. Egypt is No. 3 on the list. And at No. 4, war-torn South Sudan is expected to get even worse. Greg Stanton is a professor at George Mason University and the president of Genocide Watch. He has his own model for predicting mass killings. Stanton criticizes the Holocaust Museum's model for being overly dependent on national data that's often only released once a year.

GREG STANTON: They tend to notice that there's risk of genocide too late.

BEAUBIEN: Stanton argues that rather than looking at statistics to try to predict mass killings, you should look at events.

STANTON: It's not enough to know that you have an authoritarian regime, which is one of their characteristics. And we accept that, in fact, consider it important. But it is very important to know what that authoritarian regime is doing.

BEAUBIEN: Stanton has come up with a prediction model based on 10 stages of genocide. Interesting fact - the U.S. currently ticks off many of the early stages of a country headed for a bloodbath. There's polarization, discrimination, dehumanization. But strong legal and government institutions in the U.S. are likely to block such a disaster from happening.

The information that Genocide Watch and the Holocaust Museum are sifting through has been available to national security agencies for decades. The big question is what do you do with this information? At the time of the Rwandan genocide in 1994, Stanton was working in the State Department, and he says top government officials knew that the violence was about to begin.

STANTON: When our - President Clinton said after the Rwandan genocide, we really didn't know - I'll be direct. He was lying. He did know because I've read the cables. I've read the confidential cables that came in from Rwanda, from our ambassador there, months before that genocide. And they knew it was coming.

BEAUBIEN: Stanton's 10 stages of genocide and the Holocaust Museum's Early Warning Project are both attempts to spread information more widely about the rumblings of a genocide and get that information out as early as possible.

Jason Beaubien, NPR News.

Copyright © 2018 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.