The United States Is Not More Deadly For Civilians Than Russia
A Response to "Comparing Russian and US Airstrikes: Which are More Dangerous for Civilians?" and a Critique of the Airwars Project
Using data from the non-profit watchdog Airwars,
argues (in this piece) that American airstrikes are more deadly than Russian airstrikes for civilians (on a per strike basis). This claim piqued my interest considering Russian use of terror bombing in both Syria and Ukraine. This prompted me to take a closer look at the methodology he used but especially at that of Airwars and how the organization tallies the dead. I don’t wish for this post to be seen as an attempt to dunk on or otherwise disparage or Airwars. This post is purely a catalog of my own investigation and thoughts on the matter that I hope will spark further discussion about the way casualties are counted.
Airwars places incidents that cause civilian casualties in one of five categories: confirmed, fair, weak, contested, and discounted. In its headline estimate of civilian deaths, it combines both the “fair” and “confirmed” incidents to reach the total. This is also the number
uses for a lower bound of casualties. “Confirmed” means a belligerent has accepted responsibility for the casualties. uses Airwars’s statistics of Russian strikes in Syria and US-led strikes in Syria and Iraq as his initial point of comparison. Since Russia has acknowledged no civilian deaths in Syria, there are no confirmed deaths attributed in that category.Therefore, it is most useful to determine what Airwars considers a “fair” incident. Their description is “Where, in the view of Airwars, there is a reasonable level of public reporting of an alleged civilian casualty incident from two or more credible sources (often coupled with biographical, photographic or video evidence).” This seems a reasonable standard, but it does not distinguish between the credibility of the incident and the credibility of the attribution. It is much easier to determine that people were killed than to say who killed them.
A major limitation with using Airwars as a metric is that it does not judge proportionality. A strike that kills a hundred IS fighters and ten civilians would be undoubtedly proportionate under the laws of war. However, Airwars does not differentiate such a strike from one in which the inverse ratio is true (ten civilians are killed for one combatant). Nor does Airwars distinguish attacks that directly target civilians rather than harming them through collateral damage. The result is that civilian casualties caused by callousness or criminality are treated the same as though caused through adherence to the laws of war.
Taking an incident at random from those graded fair and attributed to the US coalition, one incident is described as the following: “‘Two unidentified men’ were reported killed in a Coalition ground operation in Al-Mananiya, east of Raqqa, on August 16th…. It is unclear at this point whether the men were civilians or belligerents.” This seems an insufficient standard by which to establish a lower bound for civilian casualties. These could be civilians. However, absent other evidence (given that these men were affirmatively targeted under the rules of engagement of the coalition) it is not justified to rate these deaths as “fair” civilian casualties. At most, they are of unknown status and (as Airwars notes) disputed by the coalition.
Let us look at another incident with the same rating, this time chosen randomly from the Russian attributed incidents in Syria:
At least two people were killed, an elderly man and an 18-year-old boy, and between three and five people were seriously injured, including a woman, a male child and a female child, when alleged Russian bombs hit a disused water station, near the villages of Urri and Ain Shib, west of Idlib, around midnight on August 22nd 2023. All of the victims were living in the water station after having been displaced and SY-24 reported they were all from the same family.
This incident, in which clear noncombatants are killed is given the same rating as an ambiguous case. Any system of grading reported casualties will have limitations, but it is clear that the range of cases within the “fair” rating mean that it cannot serve as a lower bound estimate.
In this incident, for example, some sources accuse the coalition of bombarding a hospital with white phosphorus, killing anywhere from 12-250 civilians; these allegations are flatly denied by the coalition. Furthermore, Airwars appears to count child soldiers as civilians. The page even includes the following quote: Asked to clarify the status of child soldiers according to the laws of war, Dapo Akande, Professor of Public International Law at the University of Oxford, told Airwars: “With regard to child soldiers, the fact that they are children will not change the law regarding whether they can be targeted. If they are members of an organized armed group or taking direct part in hostilities they can be targeted – even if they are doing this against the wishes of their families (or even against their own wishes).” Yet Airwars classifies this as “fair” and tallies these deaths as civilian casualties caused by coalition action. Elsewhere, IS use of human shields in assaults is also not enough to move deaths into the “contested” category of responsibility.
The methodological discrepancies do not end there. Airwars is counting Iraqi government artillery strikes to the death toll of the US-led coalition. Militia artillery strikes from the SDF and other Kurdish forces are also ascribed to the US-led coalition. The broad approach Airwars uses to attribute deaths to the coalition as a whole makes it impossible to extrapolate a general lethality of US strikes towards civilians from that data as
attempts to. It does not require any research or deep analysis to know that American airstrikes are more precise than Iraqi artillery fire. Mashing the two sets of data together tells us little that is worthwhile. You might as well say Taylor Swift and I have eleven platinum albums between us. Technically true, but not very enlightening.This is a problem that Airwars accounts for by placing attacks that cannot be identified as “contested.” Using only “confirmed” and “fair” incidents therefore omits the many attacks where Russian planes were indistinguishable from their Syrian allies. As the other list is of the “US-led coalition” in its entirety, incidents where a specific country's planes are not identifiable are not marked as contested, but instead added to the total. “Coalition” is specific enough when it comes to the US and its allies, but Airwars has decided to be more granular when it comes to Russia and Syria, leading to many more cases being marked as “contested.”
The difference in missions is also a significant explanatory factor. According to the US State Department, 90% of Russian air strikes were not targeting IS or Al-Qaeda assets in Syria, but instead moderate anti-regime groups. This is significant because if you look at many of the cases of civilian deaths in coalition strikes, the causes are related to secondary explosions from IS infrastructure located near civilians. For instance, an airstrike in support of Kurdish forces killed civilians as IS combatants had moved to a part of their camp that housed their families. In Russia’s campaigns against rebels, consisting more of conventional fighting than counter-terror, civilian casualties are more likely the product of targeting policies rather than deliberate co-location by an adversary and secondary explosions.
We need not infer in this case. There is ample evidence of Russian and Syrian government forces conducting the war crime of “double tap” strike, where a second strike is used to kill first-responders. Likewise, the Russians have demonstrably targeted civilians with incendiary munitions. Russian callousness towards civilians as a matter of organizational culture has clear precedent. The first Battle of Grozny saw as many as 25,000 civilians killed. By contrast, the Battle of Fallujah, notorious as bloody by American standards, saw 800 dead civilians.
’s use of Kharkiv as a metric for Russian lethality towards civilians is particularly unsuitable. Russian forces did not enter the city of Kharkiv itself, only the province. Perhaps most importantly, Kharkiv was a place the Russians were resoundingly defeated in relatively short order. This gave Russia little opportunity to conduct strikes that might cause mass casualties; those are uncommon when retreating in panic. Crucially, Russia has never enjoyed the air supremacy in Ukraine that it had in Syria. As such, statistics relating to Kharkiv are a poor indicator of typical Russian behavior (better exemplified by Bucha or the bombing of the theater in Mariupol, or the constant drone and missile terror bombings of Kyiv and other Ukrainian population centers).
To Kharkiv,
contrasts the Airwars statistics for deaths in Libya. However, he treats these statistics as US strikes, when they are in fact a tally of all parties’ actions. They are primarily a record of attacks resulting in civilian casualties by rival Libyan groups as well as the Turkish air force. Statistics relating to Libya from Airwars are simply not useful for a comparison of US lethality towards civilians.Conclusion
Overall, I do think the Airwars project is doing useful work. It is important to a democratic society that strikes and their effects are monitored. If I could suggest a change to the methodology, it would be to add a “proportionality” metric and a specific flag for attacks suspected of targeting civilians. Raw numbers can tell some of the story, but they omit the information necessary to judge whether coalition practices caused more casualties than were necessary to defeat IS. It is possible, however unlikely, that this was the minimum number of civilian deaths needed to defeat IS. Looking at individual cases, we can see mistakes that lead to civilian casualties. But the aggregates of Airwars do not provide us a standpoint from which to judge the overall effectiveness of the coalition’s practices.
As well, the broad category of “US-led coalition,” while an understandable limitation, presents a severe problem for using Airwars data to analyze practices. If the SDF is using inaccurate artillery pieces, that will result in high civilian losses regardless of American policy. A watchdog is most useful for critiquing military practices. A total that lumps the death tolls of militias with the USAF is not a useful tool for identifying flawed doctrine or culture.
The more fundamental issue is that these statistics, even if measured accurately, could never provide a firm basis for the kind of comparison
is seeking to make. Wars do not happen in a vacuum against an inert object. The nature of the war and the enemy will determine the civilian death toll frequently to a greater extent than the methods used in strike selection. Sheer incapacity is also a moderating force. Who can say what Russia might do to Ukraine’s cities and civilians if it could achieve air supremacy? The Russians have given us some idea in the course of their terror bombing and in their genocidal rhetoric. On the other hand, the US-led coalition fought against IS, an enemy actively seeking to instrumentalize civilians and exploit the protections they are afforded. The effects of doctrine are inevitably modified by the choices of the enemy. Use of human shields can lead to high civilian casualties even in the face of a fully cautious and moral targeting policy. Likewise, sheer weakness can mean that a policy of targeting civilians in actuality does no harm.Thus, Airwars is a bad source to actually compare the lethality of American and Russian airpower for civilians. It lumps the effects of coalition airstrikes with the casualties caused by militias and local security forces. At the same time, it fails to distinguish between deaths caused in proportional strikes, those caused by accident, and those caused by callous or deliberate targeting of civilians. The kind of statistical comparison
is attempting is simply impossible from the Airwars dataset. The misinterpretation of this data creates the false impression that Russian practices are superior in humanitarian concern to American ones or produce superior results for civilians. Neither is true, as the citizens of Kyiv, Kharkiv, Bucha, Mariupol, Grozny, and Aleppo can attest.
Hi Kiran, thank you for taking the time to read my article. Just wanted to make a few remarks.
It is fair to criticise Airwars' methodology but their flaws apply uniformly, which still makes comparing various armed parties a valid exercise. It seems that your focus on methodological flaws only applies to coalition airstrikes, not to the Russian ones.
You write "According to the US State Department, 90% of Russian air strikes were not targeting IS or Al-Qaeda assets in Syria, but instead moderate anti-regime groups." Given that the US was actively trying to overthrow the Syrian government while simultaneously arming the rebels, their designation of several groups as "moderates" is not reliable at all. David Petraeus (admittedly retired by this point) even suggested recruiting fighters from al-Nusra and ISIS to fight against both Damascus and ISIS.
https://edition.cnn.com/2015/09/01/politics/david-petraeus-al-qaeda-isis-nusra/index.html
The comparison between (the First Battle of) Grozny and Fallujah is also not particularly useful. Sure, Fallujah had far lower deaths because they emptied the city before much of the fighting. A closer comparison would be Mosul, where far more people died than in Fallujah (roughly 10,000, which is more than died in the Second Battle of Grozny) and where ten times as many civilians died than the coalition acknowledged. This is without even noting the obvious difference in post-war reconstruction in Grozny and any Iraqi city bombed by the US.
https://www.cbsnews.com/news/ap-mosul-isis-civilian-death-toll-10-times-higher-us-iraq-acknowledge/
Furthermore, if we want to expand on the humanitarian challenges due to the American fighting in Fallujah and their use of things like depleted uranium, birth defects in the city went on to be more prevalent in the Iraqi city than in Hiroshima.
https://www.aljazeera.com/features/2012/1/6/fallujah-babies-under-a-new-kind-of-siege
The claim that "it does not require any research or deep analysis to know that American airstrikes are more precise than Iraqi artillery fire" might be true but that just furthers strengthens the point that the strikes in Libya, where the US-led intervention was almost entirely from the air or sea, was far deadlier than US-led coalition in Iraq.
You are also incorrect in attributing most of the deaths in Libya to the Turkish air force and the rival groups because I am only looking at the stats from 2011 during which the Turks were far less involved. Airwars has separate data sets for Libya 2011 and Libya 2012-present. My piece only relies on the first one.
.Regarding double-tap strikes, the United States has also conducted those around the world.
https://www.thebureauinvestigates.com/stories/2013-08-01/get-the-data-the-return-of-double-tap-drone-strikes/
Finally, on the issue of proportionality, it is unclear if proportionality is particularly relevant to the question of lethality for civilians. Even if the United States is more proportional in its use of force, it can still be the case that their airstrikes are far deadlier for civilians. Given that the Ukrainian military is far bigger than the Libyan military was in 2011, it would lead to the assumption that proportionality would allow for Russian strikes to be significantly deadlier for civilians. Since we don't know what the overall combatant deaths are, we also cannot assert that the United States is more proportional in their use of force. Looking at conflicts overall, the Syrian Civil War has a lower civilian to combatant death ratio than the US war in Iraq, while the Russo-Ukrainian conflict has an even lower civilian casualty rates.
Even if we assume that the US is more proportionate (an assumption we cannot make without data to substantiate it), being a civilian in an area bombed by the US could still be more dangerous. It is remarkable how many foreign leaders regularly visit Ukraine compared to Iraq.
Airwars’ incident logs aggregate claims in local media of which they offer a rough assessment of. In most of these incidents in Syria forensic evidence is not accessible, therefore one is often not able to fully confirm the actors responsible, the number of civilian casualties, combatants killed, the nature of the target, etc. Their incident descriptions are written in a way that make this clear. As someone that does related work I’m not aware a methodologically superior way of collecting such data in such contexts. From there they and others conduct longer and more detailed investigations into incidents where stronger evidence is available. Adding a proportionality metric or some sort of checkbox for civilian vs military target would be asinine conjecture in many of these incidents. The incident description references such details with the appropriate caveats based on the sources available.