The Dark Logic Behind Armies That Rape
Leadership and strategy matter when it comes to sexual violence in warfare
In the era of #MeToo many are looking to science to better understand sexual violence—whether its why victims don’t come forward or why perpetrators commit such violence. Here I try to utilize social science to better understand why sexual violence is so pervasive during civil war.
This truly horrific problem has occurred in 64 percent of the 91 civil wars between 1980 and 2012, according to data from Harvard Kennedy School associate professor Dara Kay Cohen’s recent book, Rape During Civil War.
As I explain below, I find that militaries that dominate the state government are likely to commit widespread rape during civil wars. This finding has real policy implications for states such as the United States, which have significant influence over governments of developing nations, or humanitarian organizations trying to decide which services to bring to refugee camps.
The common explanations for rape during civil war center around religious or ethnic strife. The recent mass rape of Rohingya women by Burmese troops, and Yazidi women by ISIS fighters are just two recent examples that fit this explanation. There are other well-known mass rapes such as Bosnia and Rwanda in the 1990s.
However, the statistical tests in Cohen’s book reveal that ethnic or religious variables fail to explain mass rape.
Cohen finds, among other things, that troops with poor social ties are the most likely to commit widespread rape. When rebels or state troops are forced into fighting they must find a way to improve social ties within the group in order to survive on the battlefield.
Cohen argues that rape is a potent way of doing this, citing numerous psychological and sociological studies of gang rape. Her conclusion is also aided by months of field work involving hours of interviews with fighters who committed or witnessed rape in civil war.
At top — Chinese and Malayan girls forcibly taken from Penang by the Japanese to work as ‘comfort girls’ for the troops. Above — a Chinese girl from one of the Japanese army’s ‘comfort battalions’ sits on a stretcher, awaiting interrogation at a camp in Rangoon after World War II. Photos via Wikipedia
While Cohen’s argument is convincing she does not fully consider how battlefield conditions affect rape in civil war. And neither has any other academic or think tank. It’s true that some case studies and human rights reports point to rape as a military decision. But this argument has not been tested statistically.
Testing how the battlefield effects rape is important because a number or academics find that civilian killing in civil war—arguably a similar act to rape—is largely explained by battlefield conditions.
One camp of civilian victimization scholars focuses solely on battlefield dynamics. The first of these studies find that guerilla war itself is a strong predictor of targeting of civilians.
Similarly, another group of scholars find that good predictors of civilian victimization include wars with multiple distinct combatant groups, wars featuring long duration, battles of attrition, wars of annexation, uneven troop-to-rebel ratios and wars with high battle deaths.
A second camp of victimization scholars argue that varying types and levels of victimization correlate to variation in territorial control. For instance, Stathis Kalyvas’ seminal book, The Logic of Violence in Civil War argues that civilians living in enemy-controlled territory are likely to be killed indiscriminately. This is because the combatant is unable to gain access to good information on who is actually a threat.
Finally, although Cohen considers domestic and international politics in her study, its worth considering another scholar’s view of why these factors are important.
University of Minnesota professor Jessica Stanton’s book Violence and Restraint in Civil War focuses on domestic and international politics to explain civilian victimization.
She argues that that state leaders “use violence and restraint strategically, weighing the costs of violence based on an assessment of their own relationship with domestic and international constituents and weighing the benefits violence based on an assessment of their opponent’s relationship with its constituents.”
Domestically, she finds that unstable regimes, democracies, and ethnically inclusive regimes are likely to practice restraint from civilian targeting. Essentially, these types of regimes are fragile in the face of domestic anger or unrest. On the battlefield, these costs are reduced when rebels govern their own constituents.
Internationally, states fighting wars after the end of the cold war are most likely to practice restraint because of the heightened international costs of human rights violations during that period, according to Stanton.
Psychosocial workers with International Rescue Committee help rape survivors in South Kivu, Democratic Republic of Congo, in 2010. Photo via Wikipedia
Many of these theories provide important insight into rape in civil war. However, none of them consider the role of civil-military relations. This is surprising since civil-military relations have shown to be a vital aspect in explaining many important questions in warfare, including why militaries lose on the battlefield, and why countries select certain doctrines.
In a statistical test of a data set constructed using Stanton’s and Cohen’s data I find that governments dominated by the military are likely to implement widespread rape during civil war.
The test I use pitted my explanation against several of the variables from Cohen, Stanton, and the battlefield scholars in order see which statistically help explain widespread rape in civil war. For obvious reasons I only tested why states rape in civil war.
Cohen’s main explanation—press-ganging troops into service—failed to explain civil war rape. As did all of Stanton’s political factors, both domestic and international.
Many of the battlefield conditions variables did well. This includes Kalyvas’ control argument and Stanton’s rebel governance argument. But other factors failed to reach statistical significance, including high battle deaths, uneven troop ratios, guerrilla war and multiparty civil wars.
Borrowing from a study by Kalyvas and Laia Balcells, I also tested how civil wars fought by evenly matched opponents affects rape in civil war. These wars should, in theory, increase desperation on the battlefield. I do not find statistical significance for conventional wars. These are Syria-style civil wars in which opponents have heavy equipment and armaments.
Interestingly, however, symmetric non-conventional wars do reach statistical significance. This is likely due to the fact that these wars are fought between poorly trained, poorly disciplined troops with only light weapons. Think Sierra Leone’s civil war in the 1990s.
To come up with a measurable variable for my theory I rely on Barbara Geddes’ autocratic regimes database. This database codes various autocratic regime types over several decades, including whether a government was dominated or highly influenced by the military.
I define a government as military-dominated if the military exercised significant control of politics during a civil war or within five years prior to the start of the civil war. This latter coding decision was inspired by regimes like Burma and Indonesia where military leaders de facto rule the country even after “officially” ceding the reigns to civilians.
Again, my theory strongly and significantly correlates with widespread rape during civil war.
When left to plan and implement their own military strategy, militaries are often rigidly focused on fighting wars quickly, cheaply and with as little uncertainty as possible. For this reason, international law, ethics and even future governance of the country in the post-war era often become peripheral to the task of winning.
The grounding for this argument derives mainly from the civil-military relations literature in political science. This literature argues that military leaders usually prefer doctrines and strategies that reduce the chaos of planning for and carrying out war.
The most prominent example of this phenomenon is the predisposition of military leaders to prefer offensive doctrine. Even when doing so is clearly counterproductive. This is because planning for offensives reduces uncertainty, increasing the organizational order necessary to win.
Defensive doctrine, on the other hand, does not allow for such rigid planning. Although there are other reasons identified in the literature, fundamentally, these scholars argue that the chaos of war is the main mechanism causing military leaders to adopt strategies that reduce uncertainty and increase the chances of winning.
Interestingly, historian Isabel Hull even argues that atrocities against civilian is often a byproduct of an independent military that rigidly applies offensive doctrine.
Similarly, I argue that since rape is often used as weapon of war, such brutality often becomes widespread during war through the same institutional mechanisms that cause military leaders to prefer offensive doctrine. Indeed, rape is often used as a tool to ethnically cleanse and area, or terrorize a population into compliance with an occupying force.
Some may wonder, then, why is there little evidence of military leaders explicitly including rape as a military tactic? Following Cohen, I believe that most commanders never explicitly make such orders.
The most likely scenario is that enraged troops rape those believed to be associated with, or aiding the enemy. Other men in the unit find utility in the terror struck into the targeted community and thus join in. This occurred several times during the recent battle between Rohingya insurgents and Burmese troops.
Ground commanders also see the utility in the practice and allow it to be replicated in other targeted communities. They may explicitly order the practice, or do nothing once it begins. Of course, word of such atrocities spreads horizontally to other units, propagating the practice even further.
At some point, word of the practice eventually reaches the military leadership. If they believe there is utility of the practice they simply fail to stop the practice. These leaders may never communicate with the subordinates on the matter, but simply do nothing.
Again, military leaders may not order the practice, but when rape occurs as widely as it did in places like Burma and Bosnia, top commanders have clearly adopted this “tactic” as a part of their strategy.