- April 2012 Issue
- Targeted Killing Is Lawful If Conducted in Accordance with the Rule of Law
Targeted Killing Is Lawful If Conducted in Accordance with the Rule of Law
Posted Apr 1, 2012 4:47 AM CST
By Amos N. Guiora and Monica Hakimi
PROFESSOR GUIORA'S POSITION:
The drone attack that killed Anwar al-Awlaki has been the subject of innumerable articles, commentaries and public discussion. No doubt, the fact that al-Awlaki was an American citizen dramatically increased the public scrutiny of the drone policy initiated by President George W. Bush in the aftermath of 9/11 and significantly enhanced by President Barack Obama. The discussion is healthy and essential in large part because drone warfare will play an increasingly important role in the future of operational counterterrorism.
From the perspective of the nation-state, the benefits of targeted killing are clear: aggressive measures against identified targets with minimal, if any, risk to soldiers for the obvious reason that the killings are conducted from an unmanned aerial vehicle. But while the risks to soldiers are minimal, there are other risks that are not insignificant. Particularly, there is always the possibility of collateral damage, and there are also legitimate concerns regarding how a target is defined as legitimate.
While I believe the al-Awlaki killing was lawful, I am deeply troubled by the broad rationale articulated by the Obama administration. Yes, the al-Awlaki killing reflects aggressive self-defense coupled with a respect for the obligation to minimize collateral damage. However, the administration failed to articulate exactly how, beyond mere speech, al-Awlaki was connected to terrorist activity. The mere “likelihood” of membership in a terrorist organization is highly problematic.
The essence of targeted killing, arguably the most aggressive form of operational counterterrorism, is killing an individual the nation-state has identified as posing a danger to national security; and there is no alternative, in the name of national security, but to kill the individual. The decision must reflect a rigorous application of “checks” to ensure that the decision is neither arbitrary nor in violation of international law and core principles of morality in armed conflict.
I am a firm believer in the nation-state’s right to engage in aggressive, pre-emptive self-defense subject to powerful restraints and conditions. I advocate a measured, cautious approach to targeted killing with the understanding that the nation-state has the absolute right—and obligation—to protect its civilian population. However, that absolute right does not translate into an unlimited right.
After all, conducting operational counterterrorism divorced from a balanced approach results in violations of international law obligations and principles of morality in armed conflict; it also results in ineffective policy. The challenge in the targeted killing paradigm is to identify the specific individual deemed a legitimate target and to implement the policy in a manner reflecting respect for international law.
At its core, targeted killing reflects aggressive self-defense. Needless to say, neither the policy (in principle) nor its application (in specific) is controversy-free or immune to criticism. In the seminal case regarding targeted killings, Palestinian Society for the Protection of Human Rights and the Environment v. The Government of Israel, the Israel Supreme Court sitting as the High Court of Justice held: “The approach of customary international law applying to armed conflicts of an international nature is that civilians are protected from attacks by the army. However, that protection does not exist regarding those civilians ‘for such time as they take a direct part in hostilities’ (Section 51(3) of the First Protocol). Harming such civilians, even if the result is death, is permitted, on the condition that there is no other less harmful means, and on the condition that innocent civilians nearby are not harmed. Harm to the latter must be proportionate. That proportionality is determined according to a values-based test, intended to balance between the military advantage and the civilian damage. As we have seen, we cannot determine that a preventative strike is always legal, just as we cannot determine that it is always illegal. All depends upon the question whether the standards of customary international law regarding international armed conflict allow that preventative strike or not.”
Active self-defense in the form of targeted killing, if properly executed, not only enables the state to more effectively protect itself within a legal context but also leads to minimizing the loss of innocent civilians caught between the terrorists—who regularly violate international law by using innocents as human shields—and the state.
Active self-defense aimed at the terrorist must contain an element of “pin-pointing” so that the state will only attack those terrorists who are directly threatening society. The first step in creating an effective counterterrorism operation is analyzing the threat—including the nature of the threat, who poses it, and when it is likely to be carried out. It is crucial to assess the imminence of any threat, which significantly impacts the operational and legal choices made in response.
To ensure both the legality and morality of drone strikes, I propose the following standards: 1. A target must have made significant steps directly contributing to a planned act of terrorism. 2. An individual cannot be a legitimate target unless intelligence action indicates involvement in future acts of terrorism. 3. Before a hit is authorized, it must be determined that the individual is still involved and has not proactively disassociated from the original plan. 4. The individual’s contribution to the planned attack must extend beyond mere passive support. 5. Verbal threats alone are insufficient to categorize an individual as a legitimate target. The significant advantage of active self-defense—subject to recognized restraints of fundamental international law principles—is that the state can act against terrorists who present a real threat prior to the threat materializing (based on sound, reliable and corroborated intelligence information or sufficient criminal evidence) rather than reacting to an attack that has already occurred.
While there is much disagreement amongst legal scholars as to the meaning and, subsequently, timing of words such as planning to attack, the doctrine of active self-defense enables the state to undertake all operational measures required to protect itself.
Lawful targeted killing must be grounded in criteria-based decision-making, which increases the probability of correctly identifying and attacking the legitimate target. The state’s decision to kill a human being in the context of operational counterterrorism must be predicated on an objective determination that the “target” is, indeed, a legitimate target. Otherwise, state action is illegal, immoral and ultimately ineffective. It goes without saying that many object to the killing of a human being when less lethal alternatives are available to neutralize the target.
Any targeted killing decision must reflect consideration of four distinct elements: law, policy, morality and operational details. Traditional warfare once pitted soldier against soldier, plane against plane, tank against tank, and warship against warship. Present and future asymmetric conflict reflects state engagement with nonstate actors. In the targeted-killing paradigm, the questions of who is a legitimate target and when is the target legitimate are at the core of the decision-making process. How both questions—in principle and practice alike—are answered determines whether the policy meets international law obligations.
The dilemma of the decision-maker in the targeting paradigm is extraordinary; the time to make the decision is short, limited and stress-filled. After all, national security is at stake. However, not all individuals identified as posing a threat to national security are indeed a threat. A criteria-based decision-making model is necessary to ensure that the identified target is, indeed, the legitimate target.
Any use of force under international law must meet a four-part test: (1) it must be proportionate to the threat posed by the individual; (2) collateral damage must be minimal; (3) alternatives have been weighed, considered and deemed operationally unfeasible; and (4) military necessity justifies the action. In addition, all of these principles build on the fundamental international law principle of distinction, which requires that any attack distinguish between those who are fighting and those who are not in order to protect innocent life.
Regardless of whether a target is legitimate, if an attack fails to satisfy the requirements listed above, it will not be lawful. Thus, the Israeli Special Investigatory Commission examining the targeted killing of Salah Shehada concluded that although the targeting of Shehada—head of Hamas’ operational branch and the driving force behind many terrorist attacks—was legitimate, the extensive collateral damage caused in the attack was disproportionate.
In any targeted killing decision, three important questions must be answered: First, can the target be identified accurately and reliably? Second, does the threat the target poses justify an attack at that moment? And finally, what is the extent of the anticipated collateral damage?
To answer these questions using the criteria-based process, extensive intelligence must be gathered and thoroughly analyzed. The intelligence community receives information from three different sources: human, such as individuals who live in the community about which information is sought; signal intelligence, such as intercepted phone and email conversations; and open sources, such as the Internet and newspapers.
One of the most important questions in putting together an operational jigsaw puzzle is whether the received information is “actionable”—that is, does the information warrant a response? This question is central to the criteria-based method, or at least to a process that seeks—in real time—to create objective standards for making decisions based on imperfect information (as almost all intelligence is). It is essential that intelligence information, particularly from humans, be subjected to rigorous analysis.
Targeted killing is a legitimate and effective form of active self-defense provided that it is conducted in accordance with clear international law principles and a narrow definition of what is a legitimate target; otherwise it reflects state action bound neither by the rule of law nor constraints of morality. Morality and legality demand that operational counterterrorism measures reflect criteria-based decision-making; otherwise the stakes and the price are too high.
PROFESSOR HAKIMI'S COUNTERPROPOSAL:
I agree with much of what Professor Guiora says, but I disagree with the method he uses to get there. And I believe the method matters. Guiora assesses targeting operations under an “active self-defense” paradigm, with elements from both the jus ad bellum (the law governing the use of force) and the jus in bello (the law governing the conduct of hostilities). Under Guiora’s paradigm, a state may target terrorism suspects in anticipatory self-defense if: (1) targeting is proportional to their threat; (2) collateral damage is minimized; (3) alternatives to targeting are infeasible; and (4) military necessity justifies the action. Guiora does not explain why that paradigm is the correct one.
In fact, the ad bellum rules on defensive force probably do not govern Guiora’s poster-child case—the U.S. operation targeting Anwar al-Awlaki in Yemen. The jus ad bellum does not constrain the use of force by one state in another state where that second state consents. Yemen appears to have consented to the operation against al-Awlaki. Moreover, neither the jus ad bellum nor the traditional jus in bello requires a state to consider alternatives to lethal force—Guiora’s third criterion—if someone is a legitimate target. Finally, though Guiora argues that someone’s membership in al-Qaida is an insufficient basis for targeting him, many in bello experts treat membership in an organized armed group as dispositive. Rather than reflect existing law, then, Guiora’s model is some kind of hybrid that presents his own normative vision on when targeting should be lawful.
I assume that Guiora developed that hybrid because he believes the traditional wartime paradigm—designed for interstate wars—is poorly suited for the fight against al-Qaida. Similarly, I assume that Guiora rejects international law’s presumptive alternative—applying human rights law—because he believes that it, too, is inapposite. The human rights norms on targeting were developed for law enforcement settings. They would prohibit operations that Guiora would permit. For example, human rights law generally prohibits a state from targeting someone who is not on the verge of killing. Guiora does not require that kind of imminent threat. But Guiora does not explain why his model is preferable to the alternatives. Why should decision-makers assess targeting operations using his four criteria, instead of applying the conventional wartime paradigm, the law-enforcement paradigm of human rights law, or a hybrid advanced by someone else?
In “A Functional Approach to Targeting and Detention,” to be published this year in the Michigan Law Review, I argue that the current method for assessing targeting operations—which requires first identifying the correct legal paradigm and then applying the norms as specified for that paradigm—is misguided. The method presumes that international law’s different paradigms operate independently and sometimes incompatibly. But as I demonstrate, three core principles animate all of the international law on targeting: the jus in bello for combatants, the jus in bello for civilians and human rights law.
• The liberty-security principle identifies the outer bounds of permissible state action. The security benefits of containing someone’s threat must be proportional to or outweigh the costs of life. Targeting usually satisfies that principle if the person poses an active threat of death or serious bodily injury. In that event, the security benefit of containing the threat (protecting life or limb) is proportional to the liberty cost (taking life).
• The mitigation principle further restricts the authority to target by requiring states to try to lessen the liberty costs. States must try to contain threats using reasonable, nonlethal alternatives to targeting—most obviously, capture and detention. Reasonableness here depends primarily on two factors. One is the level of state control. The more control the state has, the more reasonable it will be to capture the suspect. The second factor is the relative efficacy of that alternative. States need not try to capture someone if doing so might compromise a security mission or fail to mitigate the liberty costs.
• The mistake principle requires states to try to verify that: (1) the person being targeted (2) poses a sufficiently serious threat (3) that cannot reasonably be contained less intrusively. In other words, the state must exercise due diligence to avoid mistakes and establish a reasonable, honest belief that its conduct is lawful. That diligence is generally less when states act in the heat of the moment than with time for deliberation. With time, states have more opportunity to ensure the accuracy of their assessments and consider the alternatives.
Those three principles govern all targeting operations but require different results depending on the facts. They may lead to results that Guiora would support. My liberty-security principle is similar to his proportionality criterion. My mitigation principle is like his criterion that states consider alternatives to targeting. And though none of his criteria specifically addresses mistakes, Guiora argues that states must gather and thoroughly analyze intelligence to ensure the accuracy of their operations. In substance, then, we seem to agree on quite a bit—at least at this level of generality. (My method does not address the permissible collateral damage. I agree with Guiora that, consistent with both the jus in bello and human rights law, any collateral damage must be minimized.)
But methodologically, we differ. I argue for assessing all targeting operations by reference to the above three principles. Most international lawyers invoke their preferred legal paradigm—Guiora selects a hybrid for “active self-defense”—and then apply the norms associated with that paradigm. As I demonstrate in my other work, that latter method breeds uncertainty and undermines the discursive process by which the law might adapt to modern challenges or hold decision-makers accountable. It breeds uncertainty because decision-makers sometimes disagree on the governing paradigm even when they agree in substance. Agreeing on the paradigm for one case may threaten slippery-slope implications for other cases. Or it may require a hybrid that, like Guiora’s, is not accessible under existing law. Debating the applicable paradigm obscures areas of substantive agreement.
More importantly, my method would focus decision-makers on the considerations that actually drive legal outcomes. Decision-makers now justify particular outcomes by invoking their preferred paradigm. Those who disagree on the applicable paradigm talk past each other, applying different norms to assess the same or similar conduct. That enfeebled discourse is a problem because international law—and especially the law on targeting—primarily operates discursively. When the legal process works well, it provides a common language with which decision-makers may justify their positions and respond to counterarguments. Eventually, they may converge on particular outcomes and resolve substantive uncertainties. But even when they disagree on substance, the discursive process helps constrain their discretion. The more persuasively an actor defends its position, the less pressure it confronts to alter its conduct. Conversely, the more compelling the counterarguments, the more it must change its behavior or refine its position to avoid condemnation.
Consider the al-Awlaki case. The U.S. government justifies that and similar operations by invoking a global armed conflict against al-Qaida. It claims that the jus in bello for traditional combatants also governs operations against members of al-Qaida. The United States makes that claim, even though it does not intend to target al-Qaida members worldwide. Rather, it claims a global conflict because it views the presumptive alternative—applying human rights law wherever U.S.-al-Qaida hostilities are not active—as sometimes too limiting. Hybrids like Guiora’s might be normatively appealing but are not now grounded in existing law. Thus, the method for assessing targeting operations pushes the United States toward a legal position that is more extreme than its practice. Meanwhile, those who protest the U.S. practice lack effective tools for holding it accountable. Just as the wartime paradigm is ineffective in legitimizing U.S. operations, the law enforcement paradigm is ineffective in constraining those operations. The United States easily dismisses human rights law as inapplicable. Of course, the al-Awlaki operation fits neatly into none of the existing paradigms. But because the current method requires identifying the correct paradigm before assessing state conduct, decision-makers endlessly debate which of those ill-fitting options is preferable.
By contrast, my method invites the United States to defend its operations on the merits, by reference to the principles that animate all existing law. Compromise positions may satisfy U.S. security needs while better legitimizing its operations internationally. The United States clearly seeks to do both. Here is President Barack Obama’s chief counterterrorism adviser, John O. Brennan, in a September address at Harvard Law School: “The effectiveness of our counterterrorism activities depends on the assistance and cooperation of our allies. ... But their participation must be consistent with their laws, including their interpretation of international law. ... The more our views and our allies’ views on these questions converge, without constraining our flexibility, the safer we will be as a country.”
Whereas the current method pushes the United States toward the extreme armed-conflict claim, mine would encourage more moderation—which the United States itself seeks. Over time, the United States and other actors may narrow their disagreements and resolve when states may target terrorism suspects extraterritorially. For example, although the United States and Human Rights Watch disagree on the applicable legal paradigm, they seem to agree that the al-Awlaki operation was permissible in part because capture was infeasible. Yet even where substantive disagreements persist, my method would better hold the United States accountable. It would require the United States to defend its conduct on the merits, instead of by reference to opaque legal paradigms. Some positions—like the claim that it may target all al-Qaida members—would be considerably more difficult to defend.
PROFESSOR GUIORA'S REBUTTAL:
To the naked eye, Professor Hakimi and I agree: Targeted killing is lawful provided it is subject to criteria and standards. Perhaps we reach this conclusion from different perspectives and distinct analyses, but the conclusion is similar. In other words, targeted killing is legal; the question is under what conditions. A close reading of Hakimi’s thoughtful and well-written response to my initial essay suggests concern with my analysis of imminence; in other words, how do we determine whether the threat posed is sufficiently imminent to determine that the potential target is, indeed, a legitimate target.
Hakimi is spot-on in highlighting this issue. Similarly, she is correct in suggesting that my essay proposes a rearticulation of international law to account for a new operational model. There is, frankly, discomfort in proposing new models; ad hoc solutions are inherently dangerous because their limits are unclear. In that vein, as history continuously suggests, unlimited executive power in the face of threats raises deeply important questions and concerns.
That said, to apply traditional models to new threats is similarly problematic; the challenge is implementing proactive operational measures subject to rigorous checks and balances with narrow definitions of critical terms. As is much discussed in scholarly literature on war and international law—and as Hakimi correctly notes—the term imminence is elusive, problematic and subject to wide interpretation. Imminence, in the targeted killing paradigm, suggests that unless the nation-state decisively engages a particular individual deemed to pose a direct threat, then innocent civilians will be harmed.
For example, to successfully conduct a suicide bombing requires a doer (the bomber), a sender (responsible for the operation in all parameters), a logistician (responsible for all operational logistics), and a financier (responsible for financing the attack, whether directly or indirectly). All four actors are essential—individually and collectively.
The proactive self-defense model at the core of targeted killing requires determining when each actor is a legitimate target predicated on an imminence analysis. Too broad a definition violates international law and morality in armed conflict standards; too narrow a definition unnecessarily endangers innocent civilians to whom the nation-state owes a duty to protect. Based on international law principles of military necessity and proportionality, along with the requirement to minimize collateral damage and to pursue alternatives, the four actors are legitimate targets at distinct times.
The doer is a legitimate target when about to commit a suicide bombing; the sender is a legitimate target 24/7 regardless of specific actions at the moment, provided collateral damage is minimized; the logistician is a legitimate target when involved in planning an attack, with the understanding that continued involvement poses a greater threat to national security than the doer of a specific attack; the financier, while largely an unresolved dilemma, is a legitimate target more akin to the sender than to the logistician—and immeasurably more so than the doer. After all, financiers are to terrorism what intelligence information is to counterterrorism. There is no terrorism without financiers, and there is no counterterrorism without intelligence information.
Where, then, does this leave us with respect to the questions Professor Hakimi posed? While recommending new paradigms is a risky proposition, the core question is whether the nation-state has the requisite tools to effectively engage in aggressive self-defense against an amorphous target. Professor Hakimi and I agree that an overbroad definition of legitimate target is a dangerous road to travel. Similarly, we agree that standardless targeted killing models not predicated on well-defined criteria pose an extraordinary danger to the rule of law and morality standards. Nevertheless, while debate is important—particularly given the dangers inherent in excessive state power—it is important to cut to the chase.
To that end, the working model proposed above for defining both the legitimate-target categories and when those targets may be legitimately engaged suggests a way forward. While inevitably subject to criticism and concern, it reflects a balancing approach required by international law in a conflict that I have previously referred to as “mission impossible.” After all, identifying a legitimate target in the traditional war paradigm posed minimal challenges to operational decision-makers; defining a legitimate target in the state/nonstate paradigm poses extraordinary challenges. Targeted killing is the most aggressive form of self-defense; in the present paradigm its morality, legality and effectiveness demand narrow definitions of the term legitimate target strictly applied. That is the model I have proposed. How criteria-based decision-making is applied determines whether the nation-state conducts itself in accordance with international law.
PROFESSOR HAKIMI'S CLOSING:
In our first round of debate, Professor Guiora proposed assessing targeting operations using a model of “active self-defense.” I presented a counterproposal. I argued that certain basic principles—which I term liberty-security, mitigation and mistake—determine when targeting is lawful.
Substantively, our proposals have much in common. They also have some differences. For example, Guiora asserts that states may target people who only finance terrorism. My liberty-security principle probably prohibits that result. Financiers do not themselves threaten bodily integrity and are too removed from the harm to justify targeting them.
No matter how one assesses those substantive differences, my proposal is preferable to Guiora’s because mine is methodologically sound. First, my proposal is rooted in existing law. Its three principles drive settled outcomes under both the jus in bello and human rights law. In contrast, Guiora’s model reflects his own normative vision. Second, my principles apply in all contexts. Guiora does not identify precisely when his model applies, but he apparently intends for it to apply in only some contexts. He did not contest my suggestion that, before using his model, decision-makers would have to determine whether it even applies. Third, and as explained in my initial response, my proposal would invigorate international law’s discursive process. Rather than debate which targeting model applies, decision-makers would focus on the considerations that actually drive legal outcomes. That substantive discourse helps develop the law and hold decision-makers accountable.
Thus far, our debate has focused on the in bello and human rights restraints on targeting. Targeting sometimes also implicates the jus ad bellum. The jus ad bellum regulates when states may use force against nonstate actors in other states. Such force is lawful when the territorial state consents. It probably also is lawful when the territorial state is unable or unwilling to contain the nonstate threat.
Since the Sept. 11 attacks, states have more frequently used force to incapacitate terrorists in other states—either with consent or under the unable-or-unwilling standard. Those operations may be lawful under the jus ad bellum without falling squarely into any in bello or human rights paradigm. For example, one-off operations might not cause sufficient violence to trigger an armed conflict. The jus in bello would not apply. Similarly, human rights law might not apply. The extent to which it applies extraterritorially is contested and uncertain. Thus, some lawyers suggest that the ad bellum license to use force effectively displaces any in bello or human rights restraints. I disagree (and I presume from Guiora’s comments that he would disagree, as well). The jus ad bellum is concerned primarily with protecting state sovereignty. The jus in bello and human rights law are about protecting individuals. Both interests are at stake when states use force extraterritorially.
In my view, decision-makers must adapt the in bello and human rights protections to account for developments under the jus ad bellum. My proposal enables that move. Recall that my mitigation principle requires states to pursue reasonable, nonlethal alternatives to targeting. Reasonableness here depends on two factors. One is state control. The greater a state’s control, the more varied its toolbox—and the more comfortably it may contain a threat without resorting to deadly force. The second factor is the relative efficacy of an alternative. States need not pursue measures that are unsuitable for or realistically might compromise the security mission.
The fact that an operation is lawful under the jus ad bellum does not make it so under my proposal. Capture might be reasonable—and therefore required—even if the territorial state consents to military force. States usually must cooperate to apprehend terrorists with the tools of law enforcement. Yet the circumstances that justify nonconsensual force might also justify taking human life. Capture might be unreasonable when the territorial state is unable or unwilling.
Consider the U.S. operation against Osama bin Laden. Pakistan did not consent to that operation. The jus ad bellum asks whether Pakistan was unable or unwilling to incapacitate bin Laden—for example, because of incompetence or corruption. If Pakistan was unable or unwilling, then working with its law enforcement apparatus was almost certainly unreasonable for purposes of the mitigation principle. To be sure, the United States might have had other alternatives for containing bin Laden’s threat. But the mitigation principle suggests that if the territorial state is uncooperative, the targeting state should have more than its ordinary, law-enforcement authorities. The United States had considerably fewer tools for controlling the situation in Abbottabad than it has domestically.
University of Utah law professor Amos N. Guiora and University of Michigan law professor Monica Hakimi discuss the rationale behind targeted killings in Patriots Debate: Contemporary Issues in National Security Law, a book sponsored by the American Bar Association’s Standing Committee on Law and National Security. The committee invited both writers.
Guiora is the son of two Holocaust survivors. He served in Israel’s Judge Advocate General’s Corps from 1986 to 2004, retiring as a lieutenant colonel. Among his posts, Guiora was legal adviser to the Gaza Strip military commander.
Hakimi, a former attorney-adviser in the Office of the Legal Adviser at the U.S. State Department, counseled on nonproliferation, the reconstruction of Iraq, international investment disputes and civil aviation. She also served as counsel before the Iran-United States Claims Tribunal and worked on cases before the International Court of Justice, U.S. federal courts and administrative agencies.
The ABA Journal’s Patriots Debate series has been publishing advance versions of the essays that will appear in the committee’s book, which is expected to be published this spring. Full versions of the essays will appear in the forthcoming book.