From the Bush to Obama to Trump administrations, the U.S. government has been engaged in a long debate with civil society organizations over the number of civilian deaths caused by U.S. counter-terrorism operations. Because large segments of the public have unrealistic expectations of the military’s ability to achieve low-to-no civilian casualties in its operations, the government finds itself in a special bind. Those public expectations have developed in part due to senior government officials over-promising. They have also developed in part due to the psychological effects of the implicit promise of precision weaponry. Nonetheless, senior military leaders like Secretary of Defense Jim Mattis are committed to avoiding and minimizing civilian casualties, and consider each instance a setback. These leaders also want to get it right in assessing the cost of warfare: They want to know how many civilians the U.S. military may have killed. The debate about the overall number of civilian casualties takes place against this backdrop.

In an opinion piece in the New York Times, I discuss a systemic defect in U.S. military procedures for assessing civilian casualties. The problem boils down to an indefensibly high standard for deciding whether an allegation of a civilian casualty is even credible enough to be investigated. What’s the standard that U.S. personnel are instructed to apply? You can find it in different places. As CENTCOM Public Affairs Officer Major Shane Huff wrote: “An allegation is determined to be credible if it can be confirmed by a preponderance of the evidence (more likely than not).” That’s the so-called “credibility determination.” In the New York Times piece, I explain why this likely results in civilian deaths left uncounted. As I wrote, “we are not grappling here with allegations of war crimes or even necessarily any wrongdoing in carrying out airstrikes.” The issue here is strictly about the correct method and results of calculating civilian casualties post-strike. Such assessments, for example, include civilian deaths that are completely unintentional, many of which resulted from ISIS’s sinister use not just of human shields but human sacrifices.

That said, there are at least two ways in which the particular defect in current civilian casualty assessments may implicate U.S. compliance with the laws of war. I thought to add to the New York Times piece, by identifying those implications here.

First, we should be concerned that the preponderance of the evidence standard for civilian casualty assessments will seep into other closely related contexts where it does not belong. One such context: U.S. military procedures for reporting incidents of law of war violations. A Department of Defense Directive requires military personnel to report and investigate incidents involving “credible information” of a potential law of war violation. But what is the standard of proof for “credible information”? If the preponderance of the evidence—which is currently used as the standard for credibility determinations for civilian casualties—were applied to reportable law of war violations, it would not only be an unnecessary public relations nightmare for the Pentagon. It would also raise serious concerns about compliance with the laws of war. While there is no legal obligation to investigate all civilian casualties in armed conflict, there is a very strong argument that States are under a legal obligation to conduct an investigation when there is a “reasonable suspicion” that an incident involved a serious violation of the laws of war—reasonable suspicion, not preponderance of the evidence. (For excellent analyses of the appropriateness of the reasonable suspicion standard in these contexts, read pieces by former Special Assistant to the U.S. Army Judge Advocate General for Law of War Matters Dick Jackson, Naval War College Professor Michael Schmitt, and Professors Amichai Cohen and Yuval Shany, to name a few.)

Second, a systematic undercounting of civilian casualties following strikes may affect the pre-strike assessment of expected civilian casualties. That’s because the military (very laudably) works to apply information about how well it performed in past strikes into its efforts to avoid and minimize civilian casualties in future strikes (see the Army’s “Civilian Casualty Mitigation Cycle”). As Larry Lewis and I recently wrote in Just Security:

Better estimating the number of civilian casualties post-strike should improve pre-strike decisions. Commanders will have a more reliable understanding of the expected damage of US operations over time. Unfortunately, this also means if DOD post-strike civilian casualty estimates are generally too low, presumably collateral damage assessments will also have generally underestimated the likelihood and magnitude of civilian casualties pre-strike (a vicious circle).

Such a defect in estimating the likelihood and magnitude of civilian casualties could create errors in any formula the military uses in its methodology for collateral damage estimates before engaging in strikes. Any systematic underestimation of civilian casualties pre-strike should also raise concerns about how well military operations are applying the laws of war (including rules on proportionality and precautions).

My piece in the Times explains the root problem. These potential legal implications raise the stakes for fixing it and not letting it spread.


Photo: Defense Secretary James N. Mattis and Marine Corps Gen. Joe Dunford, chairman of the Joint Chiefs of Staff, testify on the fiscal year 2019 defense budget request before the House Armed Services Committee in Washington, April 12, 2018. DoD photo by Navy Petty Officer 1st Class Dominique A. Pineiro