The New America Foundation uses less charitable accounting methods. But even if you adopt NAF’s high-end estimate, the drones have killed 305 civilians, compared to some 1,500 to 2,700 militants, which puts the long-term civilian death rate at about 15 percent. NAF’s figures, like LWJ’s, imply that the rate has improved: From 2010 to 2012, NAF’s high-end civilian casualty tally is 90, and its midpoint estimate of dead militants is 1,410, yielding a civilian death rate of 6 percent.
The highest reckoning of noncombatant killings comes from the Bureau of Investigative Journalism. Since 2004, BIJ counts 473 to 893 civilian deaths, against a background of roughly 2,600 to 3,500 total killings. Using BIJ’s high-end estimates, if every fatality other than a civilian is a militant, the long-term civilian death rate is 35 percent. Using BIJ’s low-end estimates, the rate is 22 percent. But again, if you break down the data by year, they point to radical improvement. From 2010 to 2012, BIJ’s count of 172 civilian deaths, against a background of 1,616 total fatalities, yields a civilian death rate of 12 percent.
In Yemen, NAF says drones have killed 646 to 928 people, of which 623 to 860 were militants. If you assume that everyone not classified as a militant was a civilian, that’s a civilian death rate of 4 to 8 percent. LWJ’s Yemen numbers are less kind: It counts 35 civilian deaths and 274 enemy deaths in 2011 and 2012, yielding a rate of 13 percent. BIJ hasn’t tallied its Yemen data, but if you add up all the fatalities it counted as civilian in 2012, you get a civilian death rate of 10 to 11 percent. (For one strike last May, which several witnesses attributed to a plane, BIJ counts more noncombatant deaths than total deaths. If you don’t include those fatalities in the drone column, the civilian death rate for 2012 is just 7 percent.)
You can argue over which of these counting systems to believe. But the takeaway is obvious: Drones kill a lower ratio of civilians to combatants than we’ve seen in any recent war. Granted, many civilians in Iraq, Afghanistan, and other such wars were killed by our enemies rather than by us. But that’s part of the equation. One reason to prefer drones is that when you send troops, fighting breaks out, and the longer the fighting goes on, the more innocent people die. Drones are like laparoscopic surgery: They minimize the entry wound and the risk of infection.
Over the years, I’ve shared many worries about the rise of drones: the illusion of withdrawal, the militarization of the CIA, the corruption of law, the evasion of congressional restraint, the risk of mission creep, and the proliferation of signature strikes. But civilian casualties? That’s not an argument against drones. It’s the best thing about them.
Correction, March 13, 2013: This article originally used the phrase "civilian casualty rate" to describe the ratio of civilian deaths to enemy combatant deaths. Casualties include injuries as well as deaths. Accordingly, the phrase has been changed to "civilian death rate" in all instances.
William Saletan's latest short takes on the news, via Twitter: