Accidental Civilian Deaths on the Rise

Charli Carpenter has earned some well-deserved attention with this post, which argues that the balance between the intentional killing of civilians and the killing of civilians as "collateral damage" has undergone a profound shift since the 19th century. The percentage of civilians killed intentionally has dropped substantially as a percentage of total civilians killed during war. This, perhaps, is not surprising, as the laws and norms that regulate war have taken an increasingly dim view of the intentional targeting of civilians. It's actually more complicated than that (rules have changed over time), but it's fair to say that a strategic bombing campaign deliberately targeting civilians, such as the United States and the United Kingdom carried out against the Axis powers in World War II, is now nearly unimaginable. The more surprising finding, however, is that the absolute number of civilians killed unintentionally (as part of other military operations) has grown substantially since 1945, and continued to grow after 1990. This is remarkable, given the steady drop in the number of inter-state wars across this period. Thus, while fewer civilians are being killed intentionally, more are being killed accidentally.

Dan Drezner notes another puzzling aspect to this trend:

Second, this bump in collateral damage also took place during a revolution in precision-guided munitions -- which, in theory, was supposed to reduce the likelihood of collateral damage.

The theory goes like this; PGMs obviate the need for saturation bombing of a particular target area by making it possible to destroy a single building or installation without destroying the rest of the neighborhood. I can think of at least two reasons why PGMs might not have produced a reduction in accidental civilian deaths. The first is that planners feel more comfortably ordering attacks in heavily populated areas when they think that they strike with precision. The collateral damage rises because of the general hubbub of civilian life (people walking past the Serbian Ministry of Intelligence when it happens to get hit), and because of PGM malfunctions and innaccuracy. The second reason is intelligence; PGM strikes depend on reliable intel about targets. The most accurate munitions in the world will produce collateral damage when they are fired at the wrong target, or at a civilian target that is inappropriately deemed hostile.

I suspect that another reason for an increase in collateral damage is casualty aversion. Casualty aversion increases the use of air attacks, and forces rules of engagement that focus more on soldier safety than on civilian survival. I wonder, though, if changes to more casualty-averse ROE increases collateral damage, or simply shifts civilian deaths from "intentional" to "collateral". For example, civilians will die in both a Soviet raid of an Afghan village in 1984, and a US raid of the same Afghan village in 2009. In the latter case, some of these civilians will die because of the inherent difficulty in distinguishing insurgent from civilian in dangerous situations, and because of the general fog of war. In the former case, civilians will die because terrorizing the village is part of the point. The civilians are dead either way (although hopefully fewer are killed in 2009), but the deaths are redefined from "intentional" to "collateral."

Indeed, this last question is relevant for evaluation of the argument as a whole. If civilian casualties are simply being redefined from "intentional" to "collateral", then the implications are somewhat different than if collateral damage is increasing independent of intentional killing. The shift could still be problematic, because redefining civilian casualties as accidental allows governments and individuals to elude legal responsibility for killing. However, it's less troubling if the overall numbers of civilian killings are on the decline.