Mass killings are commonly driven by elites who seek to consolidate power within their own group. Such atrocities serve not just to kill outgroups but also to strengthen control of the ingroup by eliminating or coercing internal opposition within the regime, as demonstrated by cases like the Rwandan genocide.

30 years ago, Rwanda experienced one of the darkest episodes in human history. Within just a few months, Hutu militias had massacred over half a million Tutsi civilians—almost two-thirds of the Tutsi population. Among Hutu, life was also cheap, however; reformist or “moderate” Hutu elites were assassinated or forced into hiding, and local Hutu officials who did not support the violence were killed. Within only two weeks the genocide had spread to all regions. By then, Hutu genocidaires had purged rival Hutu elites, and had assumed complete political control over the Hutu population—despite ultimately losing to Tutsi rebels after three and a half months of fighting.

Rwanda is not an isolated incident. There are many instances of mass killings and atrocities of innocent civilians, such as those in China, Cambodia, Indonesia, former Yugoslavia, Sudan, Myanmar, and Ethiopia. Even now, governments target civilians in large numbers.

Mass killings like those in Rwanda are especially hard to understand because they defy reason. They require both a total physical domination of an outgroup as well as the belief that this outgroup is both an imminent and existential threat. Mass killings don’t actually win wars; they generate armed resistance and invite foreign intervention. It’s striking that mass killings of outgroups are adopted by elites when they are most divided and the greatest threat to their survival comes from within their own group.

Scholars of political violence now know more about what drives mass killing than we did thirty years ago. We no longer consider them driven by age-old ethnic hatreds between peoples. We know that in Rwanda, but also in Yugoslavia and other cases of mass violence, relations between people of different ethnicities were generally cordial. Most Hutus and Tutsis had familial and friendship ties with each other. We also know that government elites ultimately drive mass killing. In many cases, extremist ideologies cause elites to perceive civilians as existential threats that cannot be negotiated with. And yet extremist ideology by itself cannot fully explain mass killings.

New research has found that mass killings are as much about threats from the ingroup as they are about perceived threats from the victim outgroup. In Rwanda, for instance, the Hutu reactionaries that drove the killing of Tutsis had their greatest competitors within the Hutu government and Hutu military.

Why do elites turn to mass killing of outgroups when the greatest threats to their political and physical survival come from within their own regime?

Evidence suggests that mass killings are often employed to help insecure elites establish control over their ingroup. First, mass killings help build coalitions with constituencies that gain from violence. This thereby builds a formidable repressive apparatus that can be used on ingroup rivals. By facilitating violence, elites provide armed thugs with the wealth, power, and status that violence provides. In such instances, mass killings can provide legitimacy, create mutual goals, and can be used as a means of paying these groups. During the Rwandan genocide, the Interahamwe militias recruited among the poorest in society that hoped to gain from the violence. These militias were useless against the well-trained Tutsi rebels, but they effectively captured local Hutu administrations in villages and towns.

Second, mass killings force a choice on local officials to facilitate or oppose the violence. In a feedback loop, the decision provides the perpetrator with information on private loyalties of officials. In contrast to coups, mass killings do not target rivals within the government directly. They therefore provide local officials with an exit option by supporting the violence. This undermines rivals at the top of the regime who are unable to mount an effective resistance.

In Rwanda, for example, the strongest resistance to the genocide was in regions in which the reformist Hutu faction was dominant. Here, local Hutu security officials successfully rallied their populations to fight against the genocide. However, as the genocide spread across communes throughout the country, reactionary Hutu “extremists” consolidated in neighbouring communes. This freed-up the militias that had been mobilised in other regions, which then made armed incursions into the reformist-controlled holdout communes. As the genocide spread, so did the pressure on local and regional security officials to step down or fall in line. Those few who didn’t were killed or were forced to flee. In only two weeks, the genocide spread from sectors and communes controlled by reactionaries to incorporate the entire Hutu state, breaking any reformist Hutu opposition in its wake.

During the final stages of this genocidal consolidation, the violence turned towards rival elites at the top of the government. Rival elites that lose their support coalitions are vulnerable and can be violently purged as traitors or collaborators with the enemy. Indeed, this purge of rival elites is a recurring feature of mass killing processes. In Rwanda, capture of the state took only two weeks: by then, reformist General Marcel Gatsinzi and the remainder of the reformist Hutu military command as well as all reformist officials in control of regional security had been purged from the regime.

In Rwanda and other cases of mass killings, the violence has been shown to be instrumental to winning elite rivalry. Mass killing, atrocity, and war are intimately related to elite rivalry in authoritarian regimes. This should induce us to think broader about political violence by these regimes. An aging Vladimir Putin, for example, likely fights his war in Ukraine in ways to maximize control over his security services, which leaves few venues for peaceful resolution. Our policies to protect civilians in Myanmar, Ethiopia, or Sudan should account for elite support coalitions. We should realise that mass killing is not simply driven by extremist leaders who desire to kill an outgroup. Autocrats kill to resolve the dangerous internal rivalries that they face.

Eelco van der Maat is Assistant Professor with tenure at Leiden University and author of the book: Elite Rivalry, Mass Killing and Genocide in Authoritarian Regimes: Why Autocrats Kill.

This article is published under a Creative Commons License and may be republished with attribution.