The Principle of Proportionality in the DoD Law of War Manual

Topline Points:

1. Feasible precautions are broader than an aspect of proportionality. Recommend not tying it to proportionality and instead discuss as a separate legal consideration.
2. Modify Section 5.12.1.3 to read as follows: “The expected loss of civilian life, injury to civilians, and damage to civilian objects is understood to mean those direct or indirect harms reasonably foreseeable resulting from the attack.”
3. Provide more relevant examples of indirect harm as two of the current examples (economic harm due to death of a family member and loss of employment due to destruction of a factory) would clearly not be considered as collateral damage under LOAC.

Introduction

International Humanitarian Law (IHL) is built upon certain core principles: military necessity, distinction, humanity, and proportionality. The principle of proportionality, along with these other core principles, is part of customary international law applicable both in international and non-international armed conflicts. The test for proportionality has been codified in Article 51(5)(b) of Additional Protocol I, which prohibits indiscriminate attacks that: “may be expected to cause incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof, which would be excessive in relation to the concrete and direct military advantage anticipated.” Additionally, the requirement to implement feasible precautions to mitigate risks to civilians in armed conflict is included with these core principles, which are universally recognized as the baseline framework for regulating armed conflict.

In July, the Department of Defense issued its latest revision to the DoD Law of War Manual. However, other areas remain where the Manual could be revised to better reflect IHL. One such possible revision would be DoD’s interpretation of the principle of proportionality with respect to: (1) the protection of military sick and wounded, military medical and religious personnel, and military medical facilities; and (2) issues of direct and indirect incidental harm. This essay will compare DoD’s interpretation of proportionality with those of the ICRC, through its Commentary to the Geneva Conventions, as well as other States’ positions through practice, reflected in part in military manuals, and offer specific suggestions for revision.

Protection of Military Sick and Wounded, Military Medical and Religious Personnel and Facilities

Section 5.10.1.2 of the DoD Law of War Manual states that proportionality “requires consideration of civilians and civilian objects, but this prohibition generally does not require consideration of military personnel and objects, even if they may not be made the object of attack, such as military medical personnel, the military wounded and sick, and military medical facilities.” The Manual’s rationale for such exclusion is two-fold. First, the treaty provisions in Additional Protocol I that prohibit attacks expected to cause excessive incidental harm do not expressly include any mention of military personnel who are protected from being made the object of attack. Second, the exclusion of military wounded and sick, and protected military personnel and military medical facilities is based on the “general impracticality of prohibiting attacks on this basis during combat operations.” (The Manual notably states the prohibition “generally” does not require consideration of these categories of individuals and objects, suggesting or leaving open the possibility that sometimes the prohibition does so require.)

Although section 5.10.1.2 of the Manual excludes consideration of military personnel and objects under the proportionality analysis, those planning attacks must take feasible precautions “to reduce the risk of harm to military personnel and objects that are protected from being made the object of attack.” The Manual reiterates this requirement in sections 7.3.3.1 (wounded and sick), 7.8.2.1 (medical and religious personnel), and 7.10.1.1 (medical facilities), setting out that although these protected persons or places do not “exempt military objectives from attack due to the risk such [personnel or objects] would be incidentally harmed, feasible precautions must be taken to reduce the risk of harm.” The Manual cites back to 5.10.1.2, which relies on Article 57 of AP I, as support for the obligation to take feasible precautions. One commentary noted that the Manual fails to explain why it appears to reject proportionality but to accept that feasible precautions must be implemented for this category of individuals and objects. This commentary argues that the rationale can be found by looking at the text of Article 57(2)(a), which not only includes protections for civilians and civilian objects, but also for people or places subject to “special protections.” This “special protections” language is not found in Article 51 of AP I.

In 2014, an ICRC legal advisor took the opposite view with respect to persons in that proportionality “must take into account possible deaths or injuries among the medical personnel, including military medical personnel, as well as combatants who are hors de combat.” In 2017, this same legal advisor expanded the argument to include military medical objects as part of the proportionality analysis. This position was adopted by the ICRC, which viewed excluding military medical personnel and objects from the proportionality analysis as “untenable” for three reasons. First, it would be incompatible with the obligations to protect military medical personnel and objects. Second, the definition of what constitutes a military objective led to the conclusion that military medical objects are considered civilian objects under IHL. Finally, the ICRC argues that the proposition that military medical personnel and objects do not enjoy protection under proportionality in attack, while civilian medical personnel and objects do, goes against the fundamental purpose of the Additional Protocols.  

This view is reiterated in the 2016 ICRC Commentary to the First Geneva Convention. In the commentary to Article 12, paragraph 1357 provides, in part, “the presence of wounded and sick members of the armed forces in the vicinity of a military objective is to be taken into consideration when carrying out a proportionality assessment prior to an attack.” This view is reiterated in the commentary to Article 19, protection of military units, in paragraph 1797 (“indiscriminate attacks affecting such establishments and units, as well as attacks that may be expected to cause excessive incidental damage to them in relation to the concrete and direct military advantage anticipated, may be deemed prohibited”), as well as Article 24, protection of personnel, in paragraph 1987 (“At a minimum, respect requires compliance with the duties of abstention, such as not to attack medical and religious personnel be it directly, indiscriminately, or in violation of the principle of proportionality”).

There is no real consensus with respect to either view. Much of the state practice and even the ICRC’s own Customary International Law Study suggests that proportionality is restricted to civilians and civilian objects. There is some state practice that can be read to include these protected persons and places as part of proportionality. For example, the New Zealand Manual notes that “even where the object of the attack is legitimate, an attack is unlawful if the destruction it is likely to cause to protected persons and objects is too great in proportion to the overall military benefit anticipated.” Similar language can be found in the Australian Manual (“Proportionality requires a commander to weigh the military value arising from the success of the operation against the possible harmful effects to protected persons and objects.”). The position of other states is less clear and leaves room for interpretation.

There is also no scholarly consensus on whether proportionality includes protected members of the armed forces. On one side, in his most recent edition of The Conduct of Hostilities Under the Law of International Armed Conflict, Yoram Dinstein concludes that the requirement of proportionality in collateral damage excludes protected members of the armed forces. On the other, Jann Kleffner, in his 2018 article Military Collaterals and Jus in Bello Proportionality, while acknowledging neither conventional nor customary law extends proportionality to protected persons, writes that the general principles of LOAC extends to a broader notion of proportionality that should also include incidental harm to protected military persons.

Issues of Direct and Indirect Incidental Harm

For a military commander to make a good faith assessment of the proportionality of an attack, he or she must calculate two things: (1) the anticipated military advantage of the attack, and (2) the expected collateral damage. This next section will deal with the latter and what a commander needs to account for when trying to determine “expected” collateral damage.

It is without dispute that incidental civilian deaths and injuries and the damage to civilian objects directly arising from the attack must be considered. However, conflicts arising in urban areas or in close proximity to military objectives and the civilian population leads to situations where attacks can cause what has been termed as “reverberating,” or “indirect” effects. In some cases, these indirect effects may be expected whereas in other cases they may only become apparent well after an attack has occurred. The issue is whether a commander must take these indirect effects into account when calculating the expected collateral damage.

Section 5.12.1.3 of the DoD Law of War Manual provides that “[t]he expected loss of civilian life, injury to civilians, and damage to civilian objects is generally understood to mean such immediate or direct harms foreseeably resulting from the attack. Remote harms that could result from the attack do not need to be considered in applying this prohibition.”

However, subsequent examples in this same section make clear that certain indirect effects are considered in this analysis. For example, attacking and destroying a power plant that would be expected to cause the loss of civilian life or injury to civilians soon after the attack due to the loss of power at a local hospital must be included in the proportionality analysis. This is consistent with Yoram Dinstein’s conclusion that “the only effects that qualify for consideration … are those that are both expected and not too remote.” This view has also been adopted in the UK Manual on the Law of Armed Conflict (“an attack on a military fuel storage depot and there is an expectation that burning fuel would flow into a residential area and cause incidental harm excessive to the anticipated military advantage”). The example in the Manual is also consistent with prior guidance such as the DoD Office of General Counsel’s Assessment of International Legal Issues in Information Operations 1999 (referencing coalition attack on electrical power system in Baghdad in Operation Desert Storm and stating “when an attack is made on an infrastructure that is being used for both military and civilian purposes the commander will not be in a proper position to weigh the proportionality of the expected military advantage against the foreseeable collateral damage unless the commander has made a reasonable effort to discover whether the system is being used for civilian purposes that are essential to public health and safety”); and the 2006 Operational Law Handbook (“[A]ttacks upon ‘dual-use’ infrastructures (i.e., those used for both military and civilian purposes) require that commanders make reasonable efforts to discover foreseeable collateral damage. Commanders must consider whether the system contemplated for attack is essential to public health and safety.”).

Modeling, such as the Collateral Damage Estimation Methodology (CDEM) plays an enormous role in assisting a commander in making a proportionality analysis. The ICRC, certain States, and commentators argue that it is insufficient to take a narrow focus of proportionality and that commanders bear a responsibility to model more than the immediate kinetic effects of an attack and should also consider how those kinetic effects will “reverberate” to produce “knock-on” effects that jeopardize the health and safety of civilians.

This was also a focus of a Chatham House report titled, Proportionality in the Conduct of Hostilities: The Incidental Harm Side of the Assessment. One of its conclusions was that incidental harm was that harm “which was reasonably foreseeable at the time the attack was launched on the basis of information that the attacker had or could reasonably be expected to have in the circumstances.” This approach, which has been adopted by the 2020 Danish Military Manual, looks to a two-part test of foreseeability of the expected harm combined with a causal link between the attack and the collateral damage.

Recommendations

This essay has briefly explored the interpretations regarding the principle of proportionality with respect to the protection of military sick and wounded, military medical and religious personnel, and military medical facilities, as well as issues related to direct and indirect civilian harm. What follows are two recommendations where the Manual could be revised with respect to these two issues.

1. Separate Feasible Precautions from Proportionality. Section 5.10.5 of the Manual notes that it “describes the requirement to take feasible precautions in planning and conducting attacks and the prohibition on attacks expected to cause excessive incidental harm as rules that are based on the principle of proportionality,” while recognizing that others describe it differently.

In my view, the requirement to take feasible precautions is broader than simply a consideration as part of the proportionality analysis and still required even when a particular attack would be proportionate. Further, feasible precautions are required as part of the principle of distinction. The Department of Defense should separate the feasible precautions section from the proportionality section. By encompassing the principle of distinction into feasible precautions, it would help clarify that military protected persons and places – including military wounded and sick, and military medical and religious personnel, and military medical facilities – are not considered military objectives and cannot be made the object of an attack and that they are accordingly protected by the obligation to take feasible precautions to minimize risk to them. Aligning the Manual with the standard understanding of the relationship between feasible precautions and proportionality (that the former is not simply a subset of the latter), could have the added benefit of clarifying the text. For example, Judge Advocates should be made aware that when the Manual, in its current version, refers to an obligation to adopt feasible precautions, it does not necessarily mean proportionality constraints.

2. Include Indirect Harm in Section 5.12.1.3 and Provide Better Examples of Indirect Harm.

a.       Section 5.12.1.3, Foreseeable Harms versus Remote Harms, provides that “[t]he expected loss of civilian life, injury to civilians, and damage to civilian objects is generally understood to mean such immediate or direct harms foreseeably resulting from the attack.” (emphasis added). This section, which currently includes only “immediate or direct” harm, should be revised to include both direct and indirect harm that is reasonably foreseeable. My recommendation to revise this language is as follows: The expected loss of civilian life, injury to civilians, and damage to civilian objects is understood to mean those direct or indirect harms reasonably foreseeable as resulting from the attack.

b.      The rest of Section 5.12.1.3 includes examples of indirect or remote harm. However, two of those examples of indirect harm (i.e., economic harm due to death of a family member and loss of employment due to destruction of a factory) are not considered collateral damage under LOAC. Accordingly, the Manual should provide more relevant examples in this section to better flesh out what types of indirect harm are reasonably foreseeable from the attack. Examples of such indirect harm would be of more value to the reader.