The Department of Defense’s third annual report to Congress on civilian casualties in U.S. military operations globally offers a few substantive additions to last year’s report, but will again leave many outside groups skeptical of department processes, and particularly skeptical of its estimates. As in past years, the report very likely understates the extent of harm caused in U.S. wars, especially in Iraq and Syria, as a result of still-unresolved shortcomings in the processes used by the U.S. military to assess and investigate harm, and a very high barrier for civilians themselves to make reports that will be taken seriously.
Even so, the unclassified report, released to Congress and the public this week and close to on time in spite of the coronavirus pandemic, provides a dose of much-needed transparency in an area that is all too often shrouded in unnecessary secrecy. The report also makes several helpful affirmative statements of policy and intent with respect to civilian casualties, provides helpful insights into current military procedures, and should be viewed in the context of several other ongoing initiatives to standardize and improve department policy (the subject of a post by our colleagues Annie Shiel and Archibald Henry forthcoming in Just Security tomorrow).
The report is required under Section 1057 of the 2018 National Defense Authorization Act (NDAA) and covers events that occurred during the prior year – 2019 in this case. This post, like those we published here in 2019 and 2018, provides a basic summary of the report, an analysis of what, if anything, is new, and a few ideas about what to expect next.
The Bottom Line and Regional Casualty Numbers
Like prior submissions, this year’s report provides the American people with a window into where the United States remains actively at war, and with whom. The report notes that it covers U.S. military operations in Iraq and Syria, Libya (specifically against ISIS but, unlike last year, not against al-Qaeda in the Islamic Maghreb), Afghanistan, Somalia (specifically against al-Shabaab), and in Yemen (specifically against al-Qaeda in the Arabian Peninsula). The report makes no mention of which groups the U.S. military has targeted in Iraq, Syria, or Afghanistan.
The report estimates that 132 civilians were killed or injured overall as a result of U.S. military operations in these locations for calendar year 2019, based on assessments as of March 2020.
As an improvement over last year’s report, and in response to the addition of the requirement in the 2020 NDAA to describe “any allegations of civilian casualties made by public or non-governmental sources formally investigated by the Department of Defense,” the Pentagon provides the number of reports and allegations it received, in addition to its own estimate of casualties for each region or named operation:
- In Iraq and Syria (Operation Inherent Resolve), out of 277 reports of incidents, the Department of Defense (DOD) estimates that 11 were “credible,” leading to an assessment of 22 civilians killed and 13 injured. By comparison, Airwars estimated that between 465 and 1,113 civilians likely died in Syria and three in Iraq from U.S.-led coalition operations, although that figure is not limited to those caused by the U.S. military. The report also makes note of an incident from 2018 that was included in last year’s report that was later attributed to “other coalition members.”
- In Afghanistan (Operations Resolute Support and Freedom’s Sentinel), DOD received 541 reports of civilian casualties, including 236 reports from NGOs and International Organizations, of which it deemed 57 “credible,” leading to an assessment of 108 civilian deaths and approximately 75 injuries. By comparison, the U.N. Assistance Mission to Afghanistan (UNAMA) estimated that in 2019, pro-government international forces (not limited to the United States) had killed 559 civilians and injured 227.
- In Somalia, the U.S. Africa Command (AFRICOM) received 27 reports of about 16 independent incidents of civilian casualties, for which it had completed assessments of 26 reports and 15 incidents. AFRICOM assessed that one incident was “credible,” which led to the death of two civilians and injury to three more. By comparison, Airwars assessed a total of 44 civilian deaths resulting from 13 separate incidents. In Libya, AFRICOM determined that one report was “not credible” after a review of “operational data, video surveillance and other data from ISR assets, and other forms of intelligence.” Airwars similarly had no reports of civilian harm from U.S. operations in Libya (but extensively documented the increase in harm from other belligerents such as the Libyan National Army, the United Arab Emirates, and Turkey.) (Note: civilian casualties reporting by AFRICOM was the recent subject of a mini-forum in Just Security.)
- The report notes that DOD received no reports of civilian casualties for Yemen, and thus assessed none as credible.
Discrepancies with Reports from Other Sources
In the NDAA (Section 1057 (b) 6), Congress required DOD to explain the general reasons for any discrepancies in its figures versus those of outside groups. DOD attributes the chasm to uneven methodological rigor applied by outside NGOs and its own access to a more complete data set, despite the ability of NGOs to gather evidence that the Defense Department is routinely either unwilling or unable to gather. This is a disappointing, but not surprising, conclusion.
The report does not consider alternative explanations that have been credibly posited, even by DOD itself, such as:
- The criteria by which the department evaluates the “credibility” of external reports presents an unnecessarily high threshold of evidence to establish that an external report is “credible” (the department itself determined that this was one reason for the discrepancy in its own internal study, as we noted previously on Just Security.).
- The definitions the department uses to evaluate the status of combatants and those participating in hostilities is different, and broader, than how those terms are interpreted by others.
- DOD’s estimates do not include any casualty assessed to have occurred as a result of the actions of its partners, which may be due to the limited scope of the congressional requirement or as a matter of its own (and its partners’) preferences and sensitivities. But this information is important for painting a more complete assessment of harm in multinational coalition operations such as Inherent Resolve and in Afghanistan, and in circumstances in which DOD personnel are actively involved in “accompany” missions with partner forces. Nor does the report, which by law is only required to cover a narrow category of information, even begin to account for the U.S. role in civilian casualties resulting from Saudi-led coalition strikes in Yemen. Congress addressed the gap in reporting on casualties in Yemen with a separate reporting requirement (Section 1274 of the FY 2020 NDAA, which required a report within 90 days of the law’s passage in December of 2019), but no such provision exists for other contexts.
Most troublingly, the reference to NGO methodology suggests that the degree to which an NGO or individual can make a claim, and expect to have that claim taken seriously, may depend on their means and capacity to employ relatively sophisticated documentation and investigative techniques. Most organizations, and certainly most individuals, could not meet these requirements. This DOD report confirms our suspicions that the overwhelming burden of proving harm rests far too greatly with those making an allegation rather than on the department conducting its own due diligence.
The additional reporting requirements added to the 2020 NDAA conveyed the sense that Congress had seized on three unresolved areas of tension: the adequacy and thoroughness of DOD’s assessments, and specifically its dismissal of NGO reports and its reluctance to interview witnesses; concern that DOD applies an overly broad definition of “combatant” to those it targets; and the lack of information about DOD’s policy for making ex gratia, or condolence, payments. Much of the language in this year’s report on these subjects remained the same as prior year reporting, with a few notable exceptions:
- This DOD report added some basic text to the basic description of the process it uses to assess external allegations, noting that engagement with NGOs and interviews are among the steps it may take to review “all readily available information” during the assessment. It does not, however, attempt to clarify whether or not the interviews to which it refers are civilian witnesses or military witnesses, nor whether civilian witnesses are understood to serve as a source of “readily available” information.
- In response to the requirement to clarify its process for making combatant status determinations, the report merely references the definitions in the Law of War Manual, rather than providing any information or clarity as to how those definitions are operationalized when assessing civilian harm. This misses the entire point of the requirement.
- The report does make two important and promising disclosures with respect to ex gratia First, DOD is soon to disclose interim policy guidance for provision of ex gratia consistent with the authorization of a $3 million fund for payments established by Section 1213 of the 2020 NDAA. The report also includes the fact that DOD made 611 ex gratia payments in 2019, which is a very significant development given that, up until this report, the widespread understanding was that ex gratia payments had largely fallen out of practice.
While they still won’t paint a complete picture, two other related reports should be made public that tell other important parts of the story on the scope and impact of the United States’ use of force around the world. One report, required under Section 1723 of the FY 2020 NDAA, must provide the overall number of strikes undertaken by the United States (as a whole, not just DOD) against terrorist targets outside what are known as “areas of active hostilities,” along with an estimate of civilian and “combatant” casualties.
This important report was also due on May 1, but because the provision did not require that the report be provided to the public, it remains unclear whether the report has been submitted. Previously, when required by Executive Order rather than statute, an unclassified version of this report had to be made public. While the statutorily required report must be submitted to Congress in unclassified form, it may include a classified annex, which would prevent congressional staff from sharing even basic statistics on the scope of U.S. lethal force operations abroad with the American people.
The second missing report, required under Section 1264 of the 2018 NDAA as amended by Section 1261 of the 2020 NDAA, is an annual report on the legal and policy framework governing the use of force and “related national security operations.” This important report, which was due on March 1, is the only comprehensive window into the executive branch’s interpretation and application of the laws and policies governing use-of-force operations and, importantly, any changes it makes to that framework.
This year, that report must also include for the first time “the criteria and any changes to the criteria for designating a foreign force, irregular force, group, or individual as lawfully targetable, as a high value target, and as formally or functionally a member of a group covered under the Authorization for Use of Military Force.” Further, the law stipulates that the unclassified portion of the report must include each change made to the legal and policy frameworks during the preceding year, a response to this administration’s practice of keeping secret such changes to previously public information.
Also coming down the pike, though farther out, is the development of DOD-wide guidance on civilian casualties prevention and response in the form of an “instruction.” In a forthcoming Just Security post, one of our colleagues will detail the recommendations that civil society organizations have made to DOD as it formulates its much-anticipated policy. Many of the recommended steps can be implemented now, without waiting for the policy to make its way through the long bureaucratic process. Ultimately, there is a lot riding on the forthcoming policy, but it remains to be seen whether it will have the necessary teeth to result in concrete improvement in DOD’s ability to predict, prevent, and respond to civilian harm. Without improving DOD’s accuracy in identifying civilians, the military will continue to be off in its pre-strike proportionality estimates as well as post-strike casualty assessments and response.
The latest civilian casualties report demonstrates how much progress DOD has made and yet how much of a paradigm shift is still needed in the department’s procedures to avoid scenarios of mistakenly targeted civilians and unaccounted-for harm that has been so painfully documented in reports like the New York Times‘ “The Uncounted.”