The U.S. DoD’s New Civilian Harm Mitigation and Response Policy and “Harm Resulting From” Military Operations
In a much-celebrated move in late December 2023, the U.S. Department of Defense (DoD) made good on its promise to publish an enterprise-wide policy on civilian harm mitigation and response (CHMR) related to U.S. military operations during armed conflict. Notably, the policy explicitly requires processes and attention to civilian protection greater than that required by the Geneva and Hague Conventions.
This emphasis follows years of Congressional interest, presidential orders, internal reviews, and public international criticism of the DoD’s alleged disregard for prioritizing pre-strike mitigation of civilian collateral damage. There was also apparent disregard for the strategic, credibility-enhancing value of candid, timely, and thorough post-strike investigations and accountability when U.S. forces caused or contributed to civilian harm. Both concerns arose and grew over the last two decades of contingency operations in the Middle East.
For its wide-viewed scope, its detailed prescriptions, and its humanitarian intentions, the new DoD policy is rightfully called a milestone and “groundbreaking” achievement. This is the first time any military has so prominently and publicly elevated civilian protection above and beyond the requirements of the laws of war as a matter of national policy, aiming to reach planners and operators at a tactical level.
DoD Instruction (DoDI) 3000.17 provides more than fifty pages of policy-level declarations such as “the protection of civilians and civilian objects is fundamentally consistent with the effective, efficient, and decisive use of force . . . and enhance[es] the effectiveness and legitimacy of military operations, and demonstrat[es] moral leadership.” It also includes detailed assignments and responsibilities, specific organizational tasks, and procedural steps and standards to actors across the DoD. Secretary of Defense Lloyd Austin, concerned with the mounting humanitarian cost in Gaza, signed the DoDI after taking a years-long interest in improving the United States’ handling of collateral damage. He has oriented the new policy toward preventing, mitigating, responding to (locally and internationally, publicly and privately), accountability for, and learning from incidents in which civilian harm “results from” U.S. military operations.
The DoDI’s attention rests in large part on new personnel at the Combatant Command level (what will become “Civilian Harm Assessment Cells”), the Combatant Commanders themselves as key decision-makers retaining much discretion over new CHMR processes, and at the Department of the Army (which will run a new Civilian Protection Center of Excellence). It does more than this of course. For a comprehensive summary of the Instruction’s many facets, Dan Stigall’s recap is an excellent start.
The DoDI follows in the wake of a broader statement of the Department’s strategic intention, the Civilian Harm Mitigation and Response – Action Plan, released in August 2022, and spearheaded by a team lead by Mr. Stigall. As the name suggests, the central focus of that plan and the new DoDI are the regular risks that military action impose on civilians stuck in and around battlefields, broadly understood. These include both human casualties and destruction or damage to civilian objects. But the Department’s new policy goes even further by opening its aperture to the proverbial second and third-order effects of those initial harms. The DoDI’s glossary defines “civilian harm” expansively to include not just unintended death or injury of civilians and destruction of their property during an otherwise lawful use of armed force, but also:
other adverse effects on the civilian population and the personnel, organizations, resources, infrastructure, essential services, and systems on which civilian life depends resulting from military operations.
This “resulting from military operations” clause is a cause for concern that seems to have largely been overlooked in CHMR policy development and public discussion. Nowhere does the new policy explain or define what it means by this term though it is integral to nearly every aspect of the DoDI’s policy.
“Civilian Harm Resulting from Military Operations”
This clause appears more than a dozen times throughout the new DoDI, including discussions of what information a Combatant Command is to make available for public view on its unclassified website and what information will be shared with Congress. It is the issue animating the anticipated updates of military doctrine and operating concepts, the central object of attention for the new Civilian Harm Assessment Cells (CHAC), and is the fundamental issue to be studied, incident-by-incident, in the new procedures for civilian harm initial reviews, assessments, and investigations.
But key to each of its goals is identifying causation. “Resulting from” is a term that signifies causation, and as such it informs subsequent designations of culpability and responsibility (if any) for that harm.
This should not be confused with an evidentiary rule like a burden of proof or persuasion. Proving causation itself is not to be understood as satisfying a burden of persuasion about facts, but about the relations between facts established by whatever burden is procedurally necessary. The DoDI establishes clearly what that burden is: as in civil law and other adversarial military administrative proceedings, the evidence must show “more likely than not” that an incident of civilian harm occurred, and “more likely than not” that the civilian harm resulted from the U.S. military during hostilities. But what individual data points or universe of facts could be logically connectable—and how they might be sufficiently related—is a question that remains unaddressed by this DoDI.
This is a problem if the DoD is serious about planning for mitigation and accountability, which it clearly is (Secretary Austin has identified it as “a strategic and moral imperative”). This causation problem is created by a lack of specificity with where causation begins and ends. Indeed, the DoDI may inadvertently complicate it even more so. Paragraph 5.2.a.(1), referring to ways in which the United States may respond to alleged civilian harm, states,
Acknowledging harm suffered by civilians and the U.S. military’s role in causing or otherwise contributing to that harm (emphasis added).
And in the section of the DoDI discussing reasons for conducting command investigations, it states,
To inquire into potential misconduct not within the purview of a Military Criminal Investigative Organization, or related matters that may have contributed to civilian harm (emphasis added).
So not only does some form of causation matter, so too does the degree to which the U.S. military “contributed” to harm, a vague term for which the DoDI includes no definition.
The new policy dictates that CHMR will be incorporated fully into deliberate planning processes (para. 3.2.a.), where legal advisors are already participating (see this doctrine manual, and this one). If commanders are to consider these “other adverse effects on the civilian population and the personnel, organizations, resources, infrastructure, essential services, and systems on which civilian life depends” that may “result from” planned military operations, this missing piece is especially concerning.
If I am the judge advocate legal counsel on that staff, part of my duties will be to ask the commander (in preparing my legal opinion) what harms or effects she foresees from a planned attack, and therefore mitigate if practicable and feasible. But will the commander simply do what she has always done, that is, use her best military judgment, experience, and insight from the rest of the staff to intuit the outer bounds of the possible civilian harm and make a common-sense assessment of cause-and-effect? If so, it is not clear that the commander would be following the prescriptions of the new CHMR policy which demand consideration of those harms that may “result from” her military operation.
Moreover, the new policy explicitly says the CHAC will have access to legal counsel during its assessments (para. 4.8.b.(3)), and investigations will be structured by regulations that also involve legal counsel (para. 4.6.b.). If the CHAC or investigating officer is to determine whether military operations resulted in certain non-obvious civilian harms, and that inquiry forms the basis for criminal or other adverse legal action, the lack of precision here may obscure the degree of the actor’s culpability and responsibility for that harm.
Imagine a scenario in which a division commander decides to use indirect fire (artillery) assets in a populated region to suppress an entrenched enemy element’s movement and to obscure the maneuver of friendly forces attempting to envelop that element. Artillery rounds from that attack impact—as foreseen—civilian homes, killing and wounding scores of civilians in the process. The subsequent initial review and civilian harm assessment—mandated by the new DoDI—conclude that it is “more likely than not” the case that this civilian harm was the result of the U.S. division commander’s decision. But that assessment and review might also determine that the collateral damage triggered a mass exodus of the now homeless civilians from that region; disease began to propagate within their internally displaced persons camp, and due to insufficient medical supplies, scores more suffered and died in the weeks that followed.
The preponderance of the evidence standard for that assessment is a far cry from the beyond a reasonable doubt standard applicable if the commander’s actions were prosecuted at a court-martial. Any “fault” of the U.S. commander for “civilian harm mitigation” purposes is unlikely to also constitute a war crime if the commander did not make those civilians the “object of the attack” (see Additional Protocol I to the 1949 Geneva Conventions, arts. 48, 51, 52), took feasible precautions, and reasonably determined that the foreseeable incidental harm to civilians was not excessive in relation to the concrete and direct military advantage he anticipated from the attack. The reasonable decision to assess and accept blame for causing the harm per the new policy, yet not prosecute the commander under such circumstances could be subject to intense public scrutiny because it would appear to be inconsistent with the CHMR’s policy emphasis on adequate preparation during planning and accountability after-the-fact. It is unlikely that a lawyer’s distinction between the two standards of proof will abate public interest or criticism.
What is Causation?
This is a psychologically, mathematically, and philosophically challenging question, and one that resides at the heart of scientific, legal, and everyday determinations of “cause-and-effect.” Scholars who consider and debate its nuances view causation in two ways or employ one of two tests: the cause-in-fact (also known as the but-for) test; and the proximate causation test. The first is the most straightforward and suited for deducing relatively simple cause-and-effect relations. For instance, would the effect have happened in the absence of the purported “cause?” If yes, then the act or omission is not the but-for cause of the effect. If a soldier shoots an unarmed and otherwise healthy prisoner of war in his custody during an interrogation, and the prisoner dies instantly following the gunshot, it is obvious what and who caused the death. But for the soldier’s unlawful attack, the prisoner would be alive. But for the artillery strike ordered by the commander in the earlier scenario, those civilians would still be alive and there would not have been a displaced persons problem at that time. This but-for reasoning is, probably, the best way to characterize most pre- and post-strike determinations of what effects will or did occur from the use of force.
But imagine a slightly more complex set of facts. Suppose a sortie of U.S. Air Force drones launches missiles on a convoy of enemy trucks traveling on an aging and poorly maintained asphalt highway. The missiles destroy the enemy vehicles and the combatants inside, but also cause significant damage to the road, which was not the intended target of the attack. The U.S. attack was the but-for cause of the incidental damage to the civilian object (the road), but what about the additional consequences? The damage to the road (pre-existing and post-strike in combination) was severe enough that its repairs took weeks, during which routine access between a nearby village and the sole clinic in the region was nearly impeded. Several health-impaired elderly villagers suffered from slowed delivery of medical care provided by a non-governmental organization’s aid workers who had been providing in-person care, one of whom eventually required airlift to a U.S. combat support hospital for a life-saving intervention.
Was that airstrike the but-for cause of that additional civilian harm, beyond the collateral road damage? If the new DoD policy demands a degree of care, consideration, and diligence above and beyond what is legally required under the law of armed conflict (LOAC), what “effect” of the airstrike should have been mitigated against during its planning? Which effect of the airstrike should the United States accept as its responsibility requiring acknowledgment, apology, or other form of condolence that the new policy emphasizes? Which harmful effect of the airstrike should be captured as a lesson-learned, incorporated into future training and planning? These same concerns will arise from the ever-increasing (lawful) use of non-kinetic cyber operations that enable or set conditions for a later or simultaneous (lawful) use of force, say, remotely shutting down portions of a city’s electrical grid.
The proximate causation test seems more appropriate for understanding a complex series of myriad harms that may arise in the context of pre-existing challenges in which a web of related and unrelated actors participates, like a planned, deliberate attack during a large-scale combat operation in an international armed conflict. But proximate causation is also notoriously difficult to explain, let alone to define and apply. The concept already finds its way into various types of military administrative investigatory processes. Army Regulation 600-8-4, Line of Duty Investigations, defines the degree of cause and effect necessary to assign a certain kind of legal responsibility for injury, illness, disease, or death of a soldier, such as when that soldier is an automobile accident driving his own car on a military installation. It notes,
The cause of an action that, in a natural and continuous sequence unbroken by an independent and unforeseeable new cause, resulting in the injury, illness, disease, or death and without which the death, injury, illness, or disease would not have been incurred.
For the purpose of assessing whether a soldier’s willful misconduct or negligence “caused” loss, damage, or destruction of government property, Department of the Army Pamphlet 735-5, paragraph 9.2.a., the Property Accountability Procedures and Financial Liability Officer’s Guide, defines proximate cause as,
the cause that, in a natural and continuous sequence, unbroken by a new cause, produces the loss or damage, and without which the loss or damage would not have occurred. Simply stated, proximate cause is the immediate or direct cause of the loss.
Superficially, these definitions are effectively synonymous (and ought to be, given that the same government body employ the same idea for the same reason, just in different contexts). Both definitions require a “natural and continuous sequence” between the cause and effect, “unbroken” by another cause, and both add what seems to be a but-for element. They observe, “without which the death . . . would not have occurred.” The first definition clarifies that if an intervening cause can be identified, and it is “independent and unforeseeable,” that intervention breaks the sequence and the original event cannot be the “proximate cause” of the undesirable consequence. The second definition, though, adds more directive “produces the loss or damage” language. Moreover, the Financial Liability regulation’s glossary inexplicably adds,
[Proximate cause] is further defined as the primary moving cause, or the predominate cause, from which the loss or damage followed as a natural, direct, and immediate consequence.
This additional language of “natural, direct, and immediate” is not synonymous with “primary moving,” and “predominate.” These terms both imply the existence of myriad facts that could be considered, if not a proximate cause, then at least a cause that contributed to the loss or damage, to one degree or another. What makes the determination of proximate causation particularly challenging in an individual case is that some facts prove that it is more likely than not an event occurred, while other facts prove causation, while others prove both, and still others prove only part of the causation definition (e.g., that a consequence was “foreseeable” but not that the effect was “immediate” or “predominate”). I previously suggested an “elements-based approach” as a more effective way for judge advocate legal advisors to communicate the nature of proximate causation, and to establish concrete analytical steps to making that determination for officers struggling through line-of-duty and financial liability investigations. It may be just as valid an approach for scoping what civilian harm does, and does not, “result from” military operations.
Concluding Thoughts
While an important development in the conduct of U.S. military operations and the reach of LOAC more generally, the new DoDI is weakened without a description of what “resulting from” means, or more specifically, what type of causation it envisions and how it would apply. The DoDI has directed the Secretary of the Army to develop a “multi-Service issuance providing supplemental guidance for conducting command-directed investigations related to incidents of civilian harm” (para. 2.14.e.). The drafters of that forthcoming guidance thus have an opportunity to determine how the concept of proximate causation will be defined and explained. It could mean the but-for view of causation, or the more nuanced proximate causation view. Or it could mean neither.
It seems reasonable to adopt a view of causation that is already the dominant view of the subject found in existing administrative investigatory regulations. If so, it is worth further consideration.
And they should. The absence of any discussion of what “resulting from” means in the context of CHMR inevitably leaves commanders and planners without enough guidance to meet the policy’s heightened expectations placed on them during the operational planning process. And no discussion of it leaves post-event investigators with insufficient guidance for assessing and assigning responsibility for causing civilian harm after military operations. This is not an insoluble problem. If the DoD can make the effort to be so explicit for mundane line-of-duty and financial liability investigations, it can and should do the same for this historic extension of legal, moral, and strategic concerns relating to civilian harm during its conduct of hostilities.
***
Lieutenant Colonel Dan Maurer is an active-duty Army Judge Advocate, presently assigned as an Associate Professor of National Security Law at The Judge Advocate General’s Legal Center and School. The personal opinions and analysis in this post are not meant to be representative of those of the U.S. government as a whole or the Army Judge Advocate General’s Corps in particular.
Photo credit: Staff Sgt. Adeline Witherspoon