Countries at Risk for Mass Killing 2021–22: Early Warning Project Statistical Risk Assessment Results

Introduction

Introduction

The Early Warning Project’s Statistical Risk Assessment uses publicly available data and statistical modeling to produce a list of countries ranked by their estimated risk of experiencing a new episode, or onset, of mass killing.

Policy makers face the challenge of simultaneously responding to ongoing mass atrocities, such as those in Burma, China, Ethiopia, South Sudan, and Syria, and trying to prevent entirely new mass atrocity situations. A critical first step toward prevention is accurate and reliable assessment of countries at risk for future violence. Earlier identification of risk broadens the scope of possible preventive actions. This report aims to help identify countries meriting preventive actions. 

In essence, our statistical model identifies patterns in historical data to answer the following question: Which countries today look most similar to countries that experienced mass killings in the past, in the year or two before those mass killings began? The historical data include basic country characteristics, as well as data on governance, war and conflict, human rights and civil liberties, and socioeconomic factors. 

This report highlights findings from our Statistical Risk Assessment for 2021–22, focusing on:

• Countries with the highest estimated risks of a new mass killing in 2021 or 2022
• Countries where estimated risk has been consistently high over multiple years
• Countries where estimated risk has increased significantly from our last assessment
• Countries with unexpected results

We recognize that this assessment is just one tool. It is meant to be a starting point for discussion and further research, not a definitive conclusion. We aim to help governments, international organizations, and nongovernmental organizations determine where to devote resources for additional analysis, policy attention, and, ultimately, preventive action. We hope that this report and our Early Warning Project as a whole inspire governments and international organizations to invest in their own early warning capabilities.

Figure 1: Heat map of estimated risk of new mass killing, 2021–22

Understanding these results

Before discussing the results, we underscore five points about interpreting this Statistical Risk Assessment:

First, as a statistical matter, mass killings are rare. On average, just over one percent of countries see a new mass killing in any given year—that means one or two countries. Our risk model predicts a similar number of new episodes of mass killing, so the average two-year risk estimate produced by our model is just under two percent. Just five out of 162 countries have a two-year risk estimate greater than ten percent, and the highest-risk country, Pakistan, is estimated to have about a one in seven chance of experiencing a new mass killing in 2021 or 2022.

Second, our model is designed to assess the risk of a new mass killing, not of the continuation or escalation of ongoing episodes. Much of the Simon-Skjodt Center’s work spotlights ongoing atrocities and urges lifesaving responses. We focus here on the risk of new mass killing to help fill an analytic gap that is critical to prevention. This feature is especially important to bear in mind when interpreting results for countries that are currently experiencing mass killings, such as Burma/Myanmar and Syria (see Figure 1 and our website for a full list of these countries). For these countries, our assessment should be understood as an estimate of the risk that a new mass killing event would be launched by a different perpetrator or targeting a different civilian group in 2021 or 2022. (Our model estimates that having a mass killing currently in progress is associated with lower risk of another one beginning.) Regardless of their ranking in this assessment, cases of ongoing atrocities demand urgent action (see Figure 4 for the Early Warning Project’s complete list of ongoing mass killings).

Third, for practical reasons, we only forecast mass killings within countries (i.e., in which the perpetrator group and the targeted civilian group reside in the same country). This risk assessment does not forecast civilian fatalities from interstate conflict. Situations in which large numbers of civilians are killed deliberately by an armed group from another country are not captured in our historical data or current forecasts. This decision does not involve a value judgment about the moral or practical significance of such atrocities, only a pragmatic judgment about what we are able to forecast reliably.

Fourth, readers should keep in mind that our model is not causal: the variables identified as predicting higher or lower risk of mass killings in a country are not necessarily the factors that drive or trigger atrocities. For example, a large population does not directly cause mass atrocities; however, countries with large populations have been more likely to experience mass killing episodes in the past, so this factor helps us identify countries at greater risk going forward. We make no effort to explain these kinds of relationships in the data; we only use them for their predictive value. An important consequence of the non-causal nature of these forecasts is that actions aimed at addressing risk factors identified in the model are not necessarily effective ways of mitigating the risk of mass atrocities; this assessment does not seek to evaluate atrocity prevention policy prescriptions. For example, although our model finds that countries coded as having severely limited freedom of movement for men are at greater risk of experiencing mass killings than are other countries, this does not imply that action to improve freedom of movement for men would help prevent mass killings. This assessment is meant to be a starting point for discussion and further research, pointing policy makers and other practitioners to the countries that merit additional analysis to determine how to prevent atrocities.

Fifth, this assessment is based on available data reflecting conditions as of the end of 2020. Events that occurred in 2021, such as the Taliban takeover in Afghanistan and the coup in Sudan, are not reflected in country risk estimates. Our assessment relies on publicly available data that is reliably measured for nearly all countries in the world, annually updated, and historically available going back many years. Because mass killing is rare, global data spanning decades are necessary to identify patterns. This means that some risk factors that might be useful predictors, but for which data meeting the above criteria are not available, are not included in the model (e.g., data on dangerous speech may be a useful predictor, but is not currently included due to a lack of data availability). Additionally, in situations where governments deliberately restrict access to international observers, such as in Burma’s Rakhine State or China’s Xinjiang Uyghur Autonomous Region, existing data might not fully reflect conditions on the ground. 

Highlights from the 2021–22 Statistical Risk Assessment 

Our model generates a single risk estimate for each country, representing the estimated risk for a new state-led or non-state-led mass killing. Figure 2 displays the estimated risk in 2021 or 2022 for the 30 highest-ranked countries. For every country in the top 30, we recommend that policy makers consider whether they are devoting sufficient attention to addressing the risks of mass atrocities occurring within that country. Strategies and tools to address atrocity risks should, of course, be tailored to each country’s context.   

Further qualitative analysis is needed to understand the specific drivers of risk in a given situation, the mass atrocity scenarios that could be deemed plausible, and the resiliencies that could potentially be bolstered to help prevent future atrocities. This kind of deeper qualitative assessment is exemplified in Early Warning Project reports on Côte d’Ivoire (2019), Mali (2018), Bangladesh (2017), and Zimbabwe (2016). Concerned governments and international organizations should consider conducting their own assessments of countries at risk,  which should suggest where adjusting plans, budgets, programs, and diplomatic strategies might help prevent mass killings in high-risk countries. Because these qualitative assessments are resource intensive, policy makers should prioritize that type of analysis on countries whose risk estimate is relatively high according to this Statistical Risk Assessment, and where opportunities for prevention exist.

Figure 2: Top 30 countries by estimated risk of new mass killing, 2021–22

Note: * Indicates ongoing state-led mass killings; ° indicates ongoing non-state-led mass killings. Some countries have multiple ongoing episodes of one or both type (e.g., Burma/Myanmar has two ongoing state-led mass killings; Nigeria has an ongoing state-led and an ongoing non-state-led mass killing). Risk-based ranking is in parenthesis. The probabilities displayed here are associated with the onset of an additional mass killing episode. See the full list of ongoing mass killings on our website. For more information on crimes against humanity in China, see later section on “Unexpected results.”

In the paragraphs below, we discuss each country’s risk according to our statistical model, and note any instances of ongoing violent conflict, group-targeted human rights abuses, and significant events that pose risk for major political instability. These brief summaries include information that goes beyond the data in our statistical model, but they are not intended to provide a comprehensive analysis of factors contributing to atrocity risk. Rather, they are intended to serve as starting points for those who are interested in deeper qualitative analysis. For each country, we also identify the specific factors that account for the risk estimates from our model (see “Methods” below for more detail on the risk factors in the model) and note whether the country is experiencing an ongoing mass killing.

---

Key questions users should ask

The results of this risk assessment should be a starting point for discussion and further analysis of opportunities for preventive action. For countries in each of the following categories, we recommend asking certain key questions to gain a fuller understanding of the risks, adequacy of policy response, and to identify additional useful lines of inquiry.

Highest-risk and consistently high-risk

• • Are the risks of large-scale, systematic attacks on civilian populations in the country receiving enough attention? 
  • What additional analysis would help shed light on the level and nature of atrocity risk in the country?
  • What kinds of crises or events (e.g., coups, elections, leadership changes, protests, etc.) might spark large-scale violence by the government or non-state actors?

Increasing risk

• • What events or changes explain the big shifts in estimated risk?
  • Have there been additional events or changes, not yet reflected in the data, which are likely to further shift the risk?
  • Is the increase part of an ongoing trend?

Unexpected results

• • What accounts for the discrepancy between the statistical results and experts’ expectations?
  • What additional analysis would help shed light on the level and nature of atrocity risk in the country? 

---

Highest-risk countries

• Pakistan (Rank: 1): Pakistan has ranked among the ten highest-risk countries each year this assessment has been produced (2014–2021); this marks the second assessment in a row that Pakistan has ranked at highest risk among all 162 countries. Pakistan faces multiple security and human rights challenges, including continued violence by the Tehrik-e-Taliban Pakistan (TTP), which claimed 26 terrorist attacks in July 2021, as well as other ideologically driven militant groups (including the Islamic State, or IS) and separatist movements. From January through September 2021, terrorist groups killed 470 people in Pakistan, including 171 civilians. The Taliban’s takeover in neighboring Afghanistan has increased concerns over the TTP’s further advancement in Pakistan. The ruling Tehreek-e-Insaf Party has continued to arrest critics. Enforced disappearances linked to the security forces remain an issue. Application of blasphemy laws, which are used to incite hatred against and attack individuals from minority religious groups, including Hindus and Ahmadis, has also expanded. According to our model, the factors accounting most for Pakistan’s high-risk estimate are its lack of freedom of movement for men, large population, high infant mortality rate, and history of mass killings. The Early Warning Project judged there was an ongoing mass killing perpetrated by the TTP and associated militias as of the end of 2020; this risk assessment relates to the possibility of a new and distinct non-state-led or state-led episode beginning, not to the ongoing episode continuing or increasing.
• India (Rank: 2): India has ranked in the top-15 highest-risk countries for the last five years, with this assessment marking its highest risk and rank to date. The Hindu nationalist-led government’s systematic discrimination against the country’s Muslim minority has intensified. In the disputed Muslim-majority territory of Jammu and Kashmir, the government ended its 18-month-long 4G Internet blackout in February 2021, though stringent anti-dissent measures remain in place. Reports of Hindu mobs attacking Muslims have continued into 2021, alongside the widespread promotion of nationalist and exclusionary ideologies. Other minorities, including Dalits, continue to face discrimination and group-targeted violence, while police-led violence against civilians and arbitrary arrests remain prevalent. Amid the COVID-19 pandemic, the government increased its restrictions on freedom of expression with wide-reaching controls over social media platforms and crackdowns on civilian protests. According to our model, the factors accounting most for India’s high-risk estimate are its large population, lack of freedom of movement for men, history of mass killing, and geographic region (South and Central Asia). The Early Warning Project judged there was an ongoing mass killing perpetrated by Naxalite-Maoists as of the end of 2020; this risk assessment relates to the possibility of a new and distinct non-state-led or state-led episode beginning, not to the ongoing episode continuing or increasing.
• Yemen (Rank: 3): Yemen has ranked sixth or higher each year since the 2017–18 risk assessment. The country’s protracted war has killed more than 12,000 civilians and given rise to approximately 20 million people in need of humanitarian assistance in 2021. The conflict intensified in early 2021 and human rights abuses persist, including the Saudi-led coalition’s indiscriminate and disproportionate use of aerial bombardment. Currently, 16.2 million people in Yemen are food insecure and more than five million are on “the brink of famine,” while warring parties have repeatedly impeded civilian access to humanitarian aid. Famine-related fatalities can count toward mass killing if they result from actions designed to compel or coerce civilian populations to change their behavior against their will, and if the perpetrators could have reasonably expected that these actions would result in widespread death among the affected populations: e.g., forced mass starvation, the intentional confiscation or destruction of healthcare supplies, forced relocation, and forced labor. According to our model, the factors accounting most for Yemen’s high-risk estimate are its lack of freedom of movement for men, its geographic region (Middle East and North Africa), that it experiences political killings that are frequently approved of or incited by top leaders of government, the presence of battle-related deaths (armed conflict between the Saudi-led coalition, Houthi rebels, and a multitude of other domestic and foreign armed actors), and its high infant mortality rate. It should be noted, as we explained in a November 2015 blog post, that our project’s definition of mass killing excludes situations in which one country’s armed forces attack civilians in another country’s territory. Thus, killings perpetrated by foreign militaries (in this case, the Saudi-led coalition) are not reflected in this forecast.

Countries in the top ten that are not discussed in this year’s report are Nigeria and Sudan. To learn more about the factors that contributed to the high-risk estimate of any of these countries, visit the country pages on our website.

Figure 3

Consistently high-risk countries

In addition to Pakistan, India, and Yemen, a few other countries have appeared near the top of our rankings for several years. 

• Afghanistan (Rank: 4): Afghanistan has ranked among the three highest-risk countries in our last five assessments and has ranked in the ten highest-risk countries since 2015. The Taliban’s August 2021 takeover following the US withdrawal and the Afghan government’s collapse has increased threats facing vulnerable groups, including women and girls and ethnic and religious communities. In particular, the Shi’a minority, who predominantly belong to the Hazara ethnic group, face a risk of crimes against humanity or even genocide, evidenced by the Taliban and other Sunni extremists’ longstanding persecution of this group and recent attacks. Factors that pose additional threats to civilians across Afghanistan include the Taliban’s apparent lack of a united or manageable command structure, the enduring al-Qaeda-Taliban alliance, and the presence of armed groups opposed to the Taliban, including the Islamic State Khorasan Province (ISKP). Despite the Taliban’s pledges to protect vulnerable groups and form an inclusive government, it has already attacked civilian protesters, appointed hardliners to top positions, and tracked down and threatened to arrest or kill groups identified as “collaborators,” among other human rights abuses. According to our model, the factors accounting most for Afghanistan’s high-risk estimate are its lack of freedom of movement for men, the presence of battle-related deaths (conflicts involving the Taliban, IS, and the Government of Afghanistan), its history of mass killing, and high infant mortality rate. The Early Warning Project judged there was an ongoing non-state-led mass killing perpetrated by the Taliban, Haqqani network, and associated armed groups as of the end of 2020; this risk assessment relates to the possibility of a new and distinct non-state-led or state-led episode beginning, not to the ongoing episode continuing or increasing. The Taliban's August takeover will be reflected in next year's risk assessment. The Simon-Skjodt Center will continue monitoring developments in Afghanistan and publicizing risks of mass atrocities against civilian groups there.
• Democratic Republic of Congo (DRC) (Rank: 5): The DRC has ranked among the ten highest-risk countries every year this assessment has been produced. With more than 130 active armed groups, violence remains high in the DRC’s eastern provinces of Ituri, North Kivu, and South Kivu. Last year, more than 800 civilians were killed by the Allied Democratic Forces (ADF) armed group in Ituri and North Kivu. Since 2019, armed group violence has displaced nearly two million people in North Kivu. In the southeastern Tanganyika province, the United Nations (UN) reported widespread and systematic gender-based sexual violence and mass displacement. Responding to increased armed group violence in May, the Congolese government imposed a “state of siege” in the eastern region, but the measure has not increased civilian security. In February 2021, the UN released a statement noting that armed group attacks on civilians “may amount to crimes against humanity,” and that Congolese security forces have violated international humanitarian law. Impunity for abuses persists, while the Tshisekedi administration has continued to commit other human rights abuses, including arbitrary arrests and harassment of critics. Recent violence has also affected the Kasai region, where a conflict resurgence between the Luba and Kuba ethnic groups has led to mass displacement, while sexual violence in the region remains widespread. According to our model, the factors accounting most for the DRC’s high-risk estimate are its lack of freedom of movement for men, large population, high infant mortality rate, and history of mass killing. The Early Warning Project judged there was an ongoing mass killing in the northeast perpetrated by various militias as of the end of 2020; this risk assessment relates to the possibility of a new and distinct non-state-led or state-led episode beginning, not to the ongoing episode continuing or increasing.
• Ethiopia (Rank: 7): Ethiopia has ranked among the top-ten highest-risk countries in our last three assessments. The civil war that began in November 2020 in Ethiopia’s Tigray region has escalated. The conflict involves Eritrean and Ethiopian federal forces, the Tigray Defense Forces (TDF), and various armed ethnic militias. In November 2021, a joint UN-Ethiopian Human Rights Commission report found that all parties may have committed crimes against humanity and war crimes. The UN had warned in February 2021 of increased risks of atrocities in the region. Reports of ethnic cleansing, gender-based sexual violence, and war crimes, among other gross violations of human rights, remain widespread. US and UN officials have denounced the use of dehumanizing speech and calls for violence by Ethiopian leaders. In August 2021, the government called on “all capable” civilians to join the war in Tigray, which has already displaced more than one-and-a-half million people. Military forces have intentionally impeded humanitarian aid access for civilians in the region, where more than five million people are in need of assistance. The government has continued to target ethnic Tigrayans with hate speech, arbitrary arrests, enforced disappearances, and other forms of violent discrimination. As the conflict spreads to other areas of the country, so too does the risk to civilians. Due to the lack of access and the ongoing conflict, independently verifiable reports of the number of people killed to date do not yet exist. As additional data become available, we may conclude that a new mass killing in Ethiopia began at the end of 2020. According to our model, the factors accounting most for Ethiopia’s high-risk estimate are its lack of freedom of movement for men, large population, history of mass killing, and the presence of battle-related deaths. The Early Warning Project judged there was an ongoing state-led mass killing against perceived state opposition in the Oromia region as of the end of 2020; this risk assessment relates to the possibility of a new and distinct non-state-led or state-led episode beginning, not to the ongoing episode continuing or increasing.
• Somalia (Rank: 11): Somalia has ranked among the top-14 highest-risk countries every year this assessment has been produced. President Mohamed Abdullahi Mohamed’s decision to extend his term amid ongoing election delays has raised concerns about increased violence. Armed conflict involving multiple clan-associated militias and state security forces continues to threaten civilian lives, while Al-Shabaab committed the majority of attacks in recent months. The electoral crisis of April 2021 risks resurgence as the president and prime minister lock heads in the lead-up to parliamentary elections in fall 2021. Analysts fear that the political dispute could distract from efforts to counter Al-Shabaab. The African Union Mission to Somalia is scheduled to transfer security duties to the Somali security forces by December 2021, furthering security concerns due to the country’s present instability. Other prevalent abuses linked to ongoing conflict in Somalia include gender-based sexual violence, widespread displacement, and violent crackdowns on journalists. According to our model, the factors accounting most for Somalia’s high-risk estimate are its lack of freedom of movement for men, high infant mortality rate, the presence of battle-related deaths (conflicts involving the Government of Somalia, Al-Shabaab, and other armed groups), and history of mass killing. The Early Warning Project judged there was an ongoing mass killing in Somalia perpetrated by Al-Shabaab as of the end of 2020; this risk assessment relates to the possibility of a new and distinct non-state-led or state-led episode beginning, not to the ongoing episode continuing or increasing.

---

Syria: The difference between new onsets and continuing mass killing

Some readers may be surprised that a country like Syria, where the scale and intensity of the war crimes and crimes against humanity are well-known, does not rank among the highest-risk countries in our assessment.

Why is Syria not ranked #1 in our risk assessment?

The percentage risk and ranking for each country represents the estimated probability that a new onset of mass killing begins in that country—that either a new perpetrator group emerges and kills more than 1,000 civilians of a specific group, or an existing perpetrator group begins targeting a new group of civilians—not that an existing mass killing continues. This decision follows the project’s goal to provide early warning before large-scale killings begin, while opportunities for prevention are greatest.

In Syria, there are two ongoing mass killings: a state-led mass killing against perceived political opposition since 2011, as well as a non-state-led mass killing perpetrated by IS and its affiliates against perceived opposition since 2012. In the case of Syria, it is difficult to imagine the state or IS targeting a new group of civilians, as the current parameters of the target groups are so broad. That means that Syria’s risk and ranking (5.6 percent risk and 14th rank) is the likelihood that a new perpetrator group emerges in 2021 or 2022.

See the Museum’s website for more information about the crisis in Syria and efforts to bring it to an end. 

---

Significant shifts in ranking

We highlight three countries that moved up in our rankings substantially between the 2020–21 and 2021–22 assessments. 

• Guinea (Rank: 6): Guinea increased significantly in our risk assessment, jumping 34 spots from 40th in 2020–21 to sixth in 2021–22. The shift can be most attributed to an increase in political killings and a decrease in freedom of movement for men, likely related to unrest around the controversial 2020 election when President Alpha Condé ran for and won what opposition parties argued was an unconstitutional third term and large-scale protests shook the country. Guinea today is in political turmoil. Citing dissatisfaction with Condé’s corrupt administration, disregard for human rights, and economic mismanagement, soldiers led by Mamady Doumbouya staged a coup on September 5, 2021, deposing Condé and appointing a new administration. Even before the coup—which will likely increase Guinea’s estimated risk when it is reflected in next year’s assessment—data indicated reason for increased concern in Guinea. According to our model, the factors accounting most for Guinea’s high-risk estimate are its lack of freedom of movement for men, high infant mortality rate, that it experiences political killings that are frequently approved of or incited by top leaders of government, its history of mass killing, and its degree of ethnic fractionalization.  
• Thailand (Rank: 19): Thailand moved up 23 spots in our risk assessment, from 42nd in 2020–21 to 19th in 2021–22. This is the first time Thailand has been ranked in the high-risk category (top-30). The shift can be most attributed to a decline in freedom of movement for men. In October 2020, the government instituted a “severe” state of emergency in response to large-scale (mostly) student-led protests. Protesters were “calling for an end to harassment of activists, abolition of Thailand’s parliament, constitutional reform, and reform of the powerful monarchy.” According to our model, the factors accounting most for Thailand’s high-risk estimate are its lack of freedom of movement for men, large population size, history of mass killing, and the presence of battle-related deaths (conflict between state security forces and the southern insurgency).
• Chad (Rank: 10): Chad moved up 13 spots in our risk assessment, from 23rd in 2020–21 to tenth in 2021–22. Chad has consistently ranked in the high-risk (top-30) category, with tenth marking its highest ranking to date. The shift can be most attributed to a decline in freedom of movement for men. In April 2020, the government issued a nationwide state of emergency in response to COVID-19, which was also used to limit freedom of assembly and target political opposition. In April 2021, President Idriss Déby, who had been in power for three decades and had just won his sixth term in a “sham” election, was killed in a battle with rebels in the country’s north. The military instantly suspended the constitution, dissolved parliament, and installed Déby’s son, Mahamat Idriss Déby Itno, as interim president and head of a Transitional Military Council to rule for 18 months until elections can be held. As of fall 2021, there is no evidence of elections being planned. Because it occurred in 2021, this coup will be reflected in the data used to generate next year’s risk assessment, and we can expect Chad’s risk to increase. In addition to national-level political instability, tensions between farmers and herders have increased intercommunal and interethnic violence in recent years. According to our model, the factors accounting most for Chad’s high-risk estimate are its lack of freedom of movement for men, high infant mortality rate, history of mass killing, the presence of battle-related deaths (conflict involving the Government of Chad, rebel groups, and IS), and its degree of ethnic fractionalization. 

Unexpected results

Global statistical risk assessments can help by identifying countries whose relatively high (or low) risk estimates surprise regional experts. In cases where our statistical results differ substantially from expectations, we recommend conducting deeper analysis and revisiting assumptions. The purpose of this analysis is not to pit qualitative analysts and statistical models against one another, but rather to deepen our understanding of risk in the country in question. We highlight three countries that, in our informal judgment, fall into this category.

• Burma/Myanmar (Rank: 17): Burma’s assessed risk may strike some as surprisingly low. In last year’s risk assessment, Burma ranked tenth. Since then, the military took power in a coup in February 2021, which has resulted in estimates of greater than 1,000 civilian deaths and apparent crimes against humanity and war crimes. Recall that this year’s risk assessment reflects data from 2020 and does not capture events that occurred in 2021, and that it is statistically rare for a country to experience multiple mass killings at a given time. If a new onset is determined in Burma in 2021, it would be the first country to experience three mass killings simultaneously. We may code a new mass killing in Burma beginning in 2021, given the military junta’s widespread violence against civilians since the coup. As discussed in a recent Simon-Skjodt Center brief, the country is experiencing an increasingly complex conflict landscape, including ethnic armed organizations and self-defense forces, an increase in Buddhist nationalist rhetoric, and warning signs of potential additional mass atrocities against the Rohingya and other ethnic and religious minorities post-coup. According to our model, the factors accounting most for Burma’s high-risk estimate are its lack of freedom of movement for men, large population, history of mass killing, and high infant mortality rate. Conversely, the fact that Burma has an ongoing mass killing is associated with lower risk. 
• China (Rank: 26): China has ranked at the edge of the top-30 “high-risk” category for the last five years. This ranking may be unexpectedly low to many observers based on the 2021 genocide determination issued by the US government and apparent crimes against humanity—including forced sterilization, torture, sexual violence, and forced labor—being perpetrated by the state against Uyghur and other Turkic Muslim populations since 2016. Due to the lack of access, independently verifiable reports of ongoing crimes in the Xinjiang Uyghur Autonomous Region are severely limited. The Chinese government is using sophisticated social and technological surveillance systems to control everyday aspects of Uyghur life and has detained “at least a million but likely closer to 3 million citizens” on the basis of their identity, the largest incarceration of an ethno-religious minority since the Holocaust. The state’s repressive campaign is multifaceted and systematic, but to date has not included widespread killing. According to our model, the factors accounting most for China’s high-risk estimate are its large population, lack of freedom of movement for men, and history of mass killing. Conversely, China’s lower-than-average infant mortality rate and the fact that political parties are banned are associated with lower risk. 
• Haiti (Rank: 60): Haiti’s ranking may strike some observers as surprisingly low, considering the major instability it is continuing to experience since the large-scale protests that began in 2019, the assassination of President Jovenel Moïse in July 2021, and continuing gang violence across the country. The UN reported increased violence in 2020, including 944 intentional homicides. Gang violence has paralyzed the country’s economic and social development, displacing thousands and exacerbating the impacts of environmental disasters (including cyclones, earthquakes, and hurricanes) on the country’s most vulnerable populations. Though the scale of violence in Haiti is consistently high, armed groups have not systematically targeted a specific group of civilians, and those responsible include multiple armed groups (gangs and state security forces) operating independently. In short, though Haiti clearly exhibits risk for violence, it does not exhibit as many characteristics that are associated with large-scale, systematic, group-targeted violence. According to our model, the factors accounting most for Haiti’s above-average risk estimate are its high infant mortality rate, history of mass killing, and that it is not a state signatory of the First Optional Protocol to the International Covenant on Civil and Political Rights. Conversely, Haiti’s lower-than-average degree of ethnic fractionalization and lack of battle deaths are associated with lower risk. 

---

A tool for assessing counterfactuals: The example of Afghanistan

The data used to produce this assessment is from 2020 (published by most sources in early- to mid-2021). This means that changes that occurred in 2021 are not captured in this risk assessment. To enable users to explore how such changes might affect a country’s risk estimate and ranking, our online platform has an interactive data tool that allows users to explore how changes to a country's risk factors would affect its risk of mass killing, holding all other variables constant. Users may want to:

1. Pose hypotheticals and assess counterfactuals (e.g., if a war were to break out in a country—captured by the “battle deaths” variable—how would that affect the risk and ranking?)
2. Manually update country risk based on known changes (e.g., knowing that a coup occurred in a country, users can see how a change in that variable affects the risk and ranking)
3. Adjust risk factor values where users disagree with a data source’s coding judgments

For example, in 2021–22, Afghanistan ranks fourth, with a 10.9 percent estimated risk. This assessment is based on 2020 data. However, someone following events in Afghanistan may suspect that events over the course of 2021—namely the fall of the Afghan government in August—may have an impact on that risk.

Using the tool, we see, for example, that if political killings become systematic and incited or approved by top government leaders, the estimate would increase from 11 percent to 17 percent risk of a new mass killing. If civil society repression increases, the estimate would increase from 11 percent to 13 percent risk of a new mass killing. If both of these variables were to change, the new risk estimate for Afghanistan would go up to 20 percent, or about a one in five chance of a new mass killing.

---

Methods

To produce this assessment, we employ data and statistical methods designed to maximize the accuracy and practical utility of the results. Our model assesses the risk for onset of both state-led and non-state-led mass killings over a two-year period.

Data

The data that inform our model come from a variety of sources. On the basis of prior empirical work and theory, we selected more than 30 variables, or risk factors, as input for our statistical model (see the discussion of our modeling approach below). All data used in our model are publicly available, regularly updated, and available without excessive delay. They also have, in our estimation, minimal risk of being retrospectively coded in ways that could depend on observed mass killings or their absence, cover all or almost all countries in the world, and go back at least to 1980 (but ideally to 1945). We include variables reflecting countries’ basic characteristics (e.g., geographic region, population); socioeconomic measures (e.g., changes in gross domestic product per capita); measures of governance (e.g., restrictions on political candidates and parties); levels of human rights (e.g., freedom of movement); and records of violent conflict (e.g., battle-related deaths, ongoing mass killings). Alongside the model, we publish a data dictionary and make the model and all data available on our GitHub repository. The only dataset the Early Warning Project maintains is that of new and ongoing mass killing events. 

Modeling approach

Our modeling approach is described in detail on our website. We use a logistic regression model with “elastic-net” regularization. In summary, based on a set of about 30 variables and data on mass killing going back to 1945, the algorithm identifies predictive relationships in the data, resulting in an estimated model. We then apply this model to recent data (from 2020 for the 2021–22 assessment) to generate forecasts. While the exact number of countries varies by year, the project includes all internationally recognized countries with populations of more than 500,000. The model automatically selects variables that are useful predictors; see our methodology page for a list of variables selected by the model. We emphasize that these risk factors should not be interpreted as causes or “drivers” of risk but simply as correlates of risk that have proven useful in forecasting.

Accuracy and uncertainty

We assessed the accuracy of this model in ways that mimicked how we use its results: We built our model on data from a period of years and then tested its accuracy on data for later years (i.e., we conducted out-of-sample testing). Our results indicate that about two out of every three countries that later experienced a new onset of mass killing ranked among the top-30 countries in a given year. See the accuracy page on our website for more details. We also analyzed the uncertainty of our model's risk rankings. This analysis gives us very high confidence, for example, that each of the 19 highest-risk countries in our 2021–22 assessment would fall within the top-30 countries, even after accounting for uncertainty due to limited data. See the uncertainty page on our website for more details.

Conclusion

Early warning is a crucial element of effective atrocity prevention. The purpose of our statistical risk assessment is to provide one practical tool to the public for assessing risk in countries worldwide. This tool should enable policy makers, civil society, and other analysts to focus attention and resources on countries at highest risk, especially those not currently receiving sufficient attention. 

This quantitative assessment is designed to serve as a starting point for additional analysis. States and international organizations have developed and implemented tools for qualitative atrocity risk assessments. We see the application of such tools as a complementary next step after our statistical analysis. These in-depth assessments should in turn spur necessary adjustments in strategic plans, budgets, programs, and diplomatic strategies toward high-risk countries. By combining these approaches—global risk assessment, in-depth country analysis, and preventive policy planning—we have the best chance of preventing future mass atrocities.

Figure 4

 

Download the report as a PDF.