By Gabriella Blum* and Philip Heymann**

Click here to download the published PDF version

I.  Introduction

Imagine that the U.S. intelligence services obtain reliable information that a known individual is plotting a terrorist attack against the United States.  The individual is outside the United States, in a country where law and order are weak and unreliable.  U.S. officials can request that country to arrest the individual, but they fear that by the time the individual is located, arrested, and extradited the terror plot would be too advanced, or would already have taken place.  It is also doubtful that the host government is either able or willing to perform the arrest.  Moreover, even if it did arrest the suspected terrorist, it might decide to release him shortly thereafter, exposing the U.S. to a renewed risk.  Should the United States be allowed to kill the suspected terrorist in the foreign territory, without first capturing, arresting, and trying him?

More than any other counterterrorism tactic, targeted killing operations display the tension between addressing terrorism as a crime and addressing it as war.  The right of a government to use deadly force against a citizen is constrained by both domestic criminal law and international human rights norms that seek to protect the individual’s right to life and liberty.  In law enforcement, individuals are punished for their individual guilt.  Guilt must be proven in a court of law, with the individual facing trial enjoying the protections of due process guarantees. Killing an individual without trial is allowed only in very limited circumstances, such as self- defense (where the person poses an immediate threat) or the immediate necessity of saving more lives.  In almost any other case, it would be clearly unlawful, tantamount to extrajudicial execution or murder.

When agents of a state seek to engage in enforcement operations outside their own territory without consent of the foreign government, they are further constrained by international norms of peaceful relations and the respect for territorial boundaries among states.  Ordinarily, when a criminal suspect finds refuge in another country, the United States would ask the other country for extradition to gain jurisdiction over him.  Even interviewing a person outside of U.S. territory would be unlawful; executing him would be an extremely egregious offense.  Violations of these norms run the risk of replacing law with force and spiraling international violence.

In wartime, governments may use deadly force against combatants of an enemy party, in which case the peacetime constraints are relaxed.  But in war, the enemy combatants belong to another identifiable party and are killed not because they are guilty, but because they are potentially lethal agents of that hostile party.  Moreover, soldiers are easily identified by the uniform they wear.  Once in the uniform of an enemy state, any soldier, by commitment and allegiance, is a potential threat and thus a legitimate target, regardless of the degree of threat the soldier is actually posing at any particular moment: the relaxing, unarmed soldier, the sleeping soldier, the retreating soldier—all are legitimate military targets and subject to intentional targeting.  No advance warning is necessary, no attempt to arrest or capture is required, and no effort to minimize casualties among enemy forces is demanded by law.

The identity and culpability of an individual not wearing a uniform but suspected of involvement in terrorism is far less easily ascertained.  While combatants should not benefit from defying the obligation to distinguish themselves from civilians (wearing civilian clothes does not give a soldier legal immunity from direct attack), the lack of uniform does raise concerns about the ability to identify individuals as belonging to a hostile force.[1] Moreover, joining a military follows a distinct procedure that allows for a bright-line rule distinguishing between those in the military and those outside it (although it hides the dangerous responsibility of civilians who take part in hostile activity without being members of the armed forces).  Joining a terrorist organization does not necessarily have a similar on/off switch; individuals might join the organization or support it in some ways or for some time, but then go back to their ordinary business without any ritual marking their joining or departing.  Identifying individuals as terrorists grows more difficult as organizations, such as Al-Qaeda, become a network of small dispersed cells, or even individuals, making the association with a hostile armed group even more tenuous.

Despite these difficulties, both the United States and Israel (as well as several other countries) have made targeted killing—the deliberate assassination of a known terrorist outside the country’s territory (even in a friendly nation’s territory), usually (but not exclusively) by an airstrike—an essential part of their counterterrorism strategy.  Both have found targeted killing an inevitable means of frustrating the activities of terrorists who are directly involved in plotting and instigating attacks from outside their territory.

Adopting a position on targeted killings involves complex legal, political, and moral judgments with very broad implications.  Targeted killing is the most coercive tactic employed in the war on terrorism.  Unlike detention or interrogation, it is not designed to capture the terrorist, monitor his or her actions, or extract information; simply put, it is designed to eliminate the terrorist.  More than any other counterterrorism practice, it reveals the complexity involved in classifying counterterrorism operations either as part of a war or as a law enforcement operation.

A targeted killing entails an entire military operation that is planned and executed against a particular, known person.  In war, there is no prohibition on the killing of a known enemy combatant; for the most part, wars are fought between anonymous soldiers, and bullets have no designated names on them.  The image of a powerful army launching a highly sophisticated guided missile from a distance, often from a Predator drone, against a specific individual driving an unarmored vehicle or walking down the street starkly illustrates the difference between counterinsurgency operations and the traditional war paradigm.  Moreover, the fact that all targeted killing operations in combating terrorism are directed against particular individuals makes the tactic more reminiscent of a law enforcement paradigm, where power is employed on the basis of individual guilt rather than status (civilian/combatant).  Unlike a law enforcement operation, however, there are no due process guarantees: the individual is not forewarned about the operation, is not given a chance to defend his innocence, and there is no assessment of his guilt by any impartial body.

The uneasiness about classifying and evaluating targeted killings further grows as these operations are carried out outside an immediate battlefield, such as in Yemen, Pakistan, or Somalia.  Justifying targeted killings in those countries faces the challenges of the constraints of peaceful international relations or else a potentially unlimited expansion of the geographical scope of the armed conflict beyond the immediate theater of war.  There are slippery slope concerns of excessive use of targeted killings against individuals or in territories that are harder to justify.  Recent reports about a U.S. “hit list” of Afghan drug lords, even though supposedly taking place in an active combat zone, have sparked criticism that drug lords, even when they finance the Taliban, do not fit neatly within the concept of “combatant,” and must instead be treated with law enforcement tools.[2]

Concerns about the use of targeted killings grow as collateral harm is inflicted on innocent bystanders in the course of attacks aimed at terrorists.  In war, collateral damage to civilians, if proportionate to the military gain, is a legitimate, however dire, consequence of war.  In domestic law enforcement, the police must hold their fire if they believe that there is a danger to innocent bystanders, except where using lethal force against a suspect is reasonably believed likely to reduce the number of innocent deaths.

To make this tactic acceptable to other nations, targeted killings must be justified and accounted for under a set of norms that may not correspond perfectly to either peacetime or wartime paradigms, but is nonetheless respectful of the values and considerations espoused by both.  In this chapter we consider the advantages and disadvantages of choosing either paradigm as our starting point, thereafter subjecting the paradigm to necessary modifications for application to the counterterrorism context.  We do so by assessing the American and Israeli experience in employing targeted killings and its legal, moral, and strategic implications.

II.  The Practice of Targeted Killing

A.  The United States

Countries have been in the business of targeted assassinations for centuries.  The United States has been a more recent participant.  The U.S. Senate Select Committee chaired by Senator Frank Church (the Church Committee) reported in 1975 that it had found evidence of no less than eight plots involving CIA efforts to assassinate Fidel Castro, as well as assassination plots against President Ngo Dinh Diem of South Vietnam and General Rene Schneider of Chile.  During the Vietnam War, the Phoenix Program planned the assassination of Viet Cong leaders and sympathizers.  In 1986, President Ronald Reagan ordered Operation El Dorado Canyon, which included an air raid on the residence of Libyan ruler Muammar Qaddafi.  Qaddafi remained unscathed, but his daughter was killed.

Assassination plots by both the United States and other countries were not publicly acknowledged, justified, or accounted for.  Rather, they were taken to be an element of that part of foreign relations that always remains in the dark, outside official protocol or lawful interaction, unspoken of, but understood to be “part of the international game.”  Many of the plots never became public knowledge; few, if any, enjoyed enduring public acceptance.

The political fallout of the Church Committee’s criticism of the covert assassination program during the Cold War brought President Gerald Ford to promulgate an executive order banning assassinations, a prohibition that was later incorporated into Executive Order 12333 (1981) signed by President Ronald Reagan and that remains in effect today.  The executive order was part of the reason that those responsible for planning military actions prior to 1998 took great care to avoid any appearance of targeting specific individuals.

However, following the 1998 bombings of the American embassies in Kenya and Tanzania, and on the basis of a (secret) favorable legal opinion, President Bill Clinton issued a presidential finding (equivalent to an executive order) authorizing the use of lethal force in self-defense against Al-Qaeda in Afghanistan.  Shortly thereafter, seventy-five Tomahawk cruise missiles were launched at a site in Afghanistan where Osama Bin Laden was expected to attend a summit meeting.  Following the attacks of September 11, 2001, President George Bush reportedly made another finding that broadened the class of potential targets beyond the top leaders of Al-Qaeda, and also beyond the boundaries of Afghanistan.  Secretary of Defense Donald Rumsfeld ordered Special Operations units to prepare a plan for “hunter killer teams,” with the purpose of killing, not capturing, terrorist suspects.  Using the war paradigm for counterterrorism enabled government lawyers to distinguish lethal attacks on terrorists from prohibited assassinations and justify them as lawful battlefield operations against enemy combatants, much like the uncontroversial targeted killing of Japanese Admiral Isoroku Yamamoto while he was traveling by a military airplane during World War II.  According to reports, President Bush also gave the CIA, and later the military, authority to kill U.S. citizens abroad if there was strong evidence that an American was involved in organizing or carrying out acts of terrorism against the United States or U.S. interests.[3]

The first publicly known targeted killing of terrorists outside a theater of active war under the most recent presidential finding was in Yemen in November 2002, when a Predator (unmanned and remotely operated) drone was launched at a car carrying Al-Harethi, suspected of the USS Cole bombing, along with four others, one of whom was an American citizen.  The attack in Yemen was executed with the approval of the government of Yemen, thereby eliminating some of the international legal difficulties associated with employing force in another country’s territory.

Later, the United States engaged in a number of targeted killing operations in Pakistan, not all of which were authorized or approved by the Pakistani government.  One of those operations, carried out in January 2006 and directed at Bin Laden’s deputy, Aiman al-Zawahiri, left eighteen civilians dead while missing al-Zawahiri altogether and drawing fierce domestic criticism of then-Pakistani President Pervez Musharraf.

Since 9/11, Predator drones have reportedly been used dozens of times by the United States to fire on targets in Afghanistan, Iraq, Pakistan, Yemen, and elsewhere.  The targeted killing operations have successfully killed a number of senior Al-Qaeda members, including its chief of military operations, Mohammad Atef.

President Barack Obama’s administration has not changed the policy on targeted killings; in fact, it ordered a “dramatic increase” in the drone-launched missile strikes against Al-Qaeda and Taliban members in Pakistan.  According to commentators, there were more such strikes in the first year of Obama’s administration than in the last three years of the Bush administration.  CIA operatives have reportedly been involved in targeted killing operations in Yemen and Somalia as well, although in Yemen the operations are carried out by Yemeni forces, with the CIA assisting in planning, munitions supply, and tactical guidance.  Obama has also left intact the authority granted by his predecessor to the CIA and the military to kill American citizens abroad, if they are involved in terrorism against the United States.[4]

B.  Israel

Since its creation in 1948, Israel has assassinated various enemy targets, including Egyptian intelligence officers involved in orchestrating infiltrations into Israel in the 1950’s, German scientists developing missiles for Nasser’s Egypt in the 1960’s, Black September members following the Munich Olympics massacre of 1972, and prominent leaders of Palestinian and Lebanese terrorist networks such as the secretary general of Hezbollah in 1992.  Israel even planned an assassination operation against Saddam Hussein after the Gulf War.

But it was only during the Second Intifada, which began in September 2000, that targeted killings became a declared and overt policy in the fight against terrorism.  Since the first publicly acknowledged targeted killing operation by Israel in November 2000, there have been many dozens of such operations, mostly in Gaza and only rarely in the West Bank.  The use of targeted killing operations increased with the level of Palestinian violence and decreased with the prospects of peaceful relations between the parties.  Following waves of suicide bombings, there was a surge in targeted killing operations; when there were declarations of ceasefire or when political processes were underway, operations were halted.

The process for approving targeted killing operations in Israel involves an intelligence “incrimination” of the target, which identifies the target as a person actively involved in acts of terrorism; a plan for the time, place, and means of the attack (most commonly, an airstrike); consideration of the danger of collateral damage; and a review of potential political ramifications.  The complete plan must receive the approval of a top-level political official.  There is no external review process, judicial or other.

The stated Israeli policy is that only members of a terrorist organization who are actively involved in an ongoing and direct manner in launching, planning, preparing, or executing terrorist attacks are lawful targets.  In addition, targeted killing operations will not be carried out where there is a reasonable possibility of capturing the terrorist alive.

The legitimacy and usefulness of the practice of targeted killings has been hotly debated within Israel ever since it became publicly known that Israel was employing them.  No incident illustrates the tension between the benefits of a legitimate procedure meeting due process standards and the national security demands for exigency better than the targeted killing of Salah Shehadeh.  Shehadeh was the head of the military wing of Hamas in the Gaza Strip, and was, according to Israeli intelligence, directly responsible for the killing of scores of Israeli civilians and soldiers and the injury of hundreds of others in dozens of attacks.

Initially, Israeli officials had demanded that the Palestinian Authority arrest Shehadeh.  When the Palestinian Authority declined, the Israeli government sought to capture him directly, but had to forego such plans when it realized that Shehadeh lived in the middle of Gaza City, where no Israeli soldiers had been deployed since 1994, and where any attempt to apprehend Shehadeh would turn into a deadly confrontation.  It was then that Israel decided to kill him.

On the night of July 22, 2002, an Israeli F-16 aircraft dropped a single one-ton bomb on Shehadeh’s house in a residential neighborhood of Gaza City, one of the most densely populated areas on the globe.  As a result, Shehadeh and his aide, as well as Shehadeh’s wife, three of his children, and eleven other civilians, most of whom were children, were killed.  One hundred and fifty people were injured.

Israeli officials claimed that the targeted killing of Shehadeh was designed to prevent him from carrying out future attacks against Israelis.  They asserted that, according to intelligence reports, at the time of his killing, Shehadeh was effectively a “ticking bomb,” in the midst of planning at least six different attacks on Israelis, including one designed as a “mega-attack,” involving a truck loaded with a ton of explosives.

In the aftermath of the attack, there was little disagreement that Shehadeh himself was a justified target.  Nonetheless, television images of funerals of slain children drew fierce criticism both within and outside of Israel.  Legal proceedings were initiated in Britain against the Israel Defense Force’s (IDF) Chief of General Staff, the IDF’s air force commander, and the commander of the Southern Command.[5] A lawsuit under the Alien Tort Claims Statute and the Torture Victim Protection Act was filed by the Center for Constitutional Rights in the Southern District of New York against the head of the Israel Security Agency at the time, Avi Dichter.[6] The claim was subsequently dismissed by the court.[7]

Within Israel, the cars of air force pilots, normally considered demigods in popular Israeli culture, were sprayed with graffiti insults of “war criminal.”  A year later, twenty-seven pilots declared that they would refuse to carry out any additional bombing missions in Gaza.  Israeli leftwing activists petitioned the High Court of Justice to order a criminal investigation into the attack and also to prevent the air force commander—Major General Dan Halutz—from being promoted to Deputy Chief of General Staff (Halutz later became Chief of General Staff, but resigned after the 2006 war in Lebanon).  A criminal proceeding was initiated in Spain by relatives of the victims of the attack on Shehadeh against seven Israeli officials for alleged war crimes (and was later dismissed by a Spanish court).

In a traditional war context, killing fourteen civilians along with the highest military commander of the enemy could be considered proportionate collateral damage.  For comparison’s sake, the special report of the prosecutor of the International Criminal Tribunal for the Former Yugoslavia on the NATO operation in Kosovo determined that ten (and according to some reports, seventeen) civilian casualties were legitimate collateral damage for the attack on the Serbian television station.

But public opinion could not disentangle the proportionality question from the broader political context of the Israeli-Palestinian relationship: the legality and morality of the continued occupation of Gaza and the West Bank (Israel withdrew from Gaza three years later); the perception of failure in conducting the war on terrorism; and the frustration over losing the symbolic struggle over “victimhood” to the Palestinians.

A year after the targeted killing of Shehadeh, ten senior Hamas leaders met in a room on the top floor of a residential building in Gaza.  Bruised by the effects of the Shehadeh operation, the Israeli security agencies decided to use a laser-guided bomb only a quarter of the size of the one used to kill Shehadeh.  The Hamas leaders left the room seconds before the bomb hit.  The top floor was destroyed, but the group escaped with minor scratches.  Had a larger bomb been used, the building would have collapsed, together with the Hamas leadership and civilian residents.

Two years later, in a newspaper interview, Avi Dichter, while admitting that the pre-operation assessment misjudged the level of collateral damage that would result from the attack on Shehadeh, added that “he couldn’t say how many Israelis paid with their lives for the fact that Shehadeh continued to operate long after Israel had the operational capability to harm him, but not the moral will to do it.”  In describing the subsequent failed attack on the Hamas leadership as “a miss,” Dichter lamented, “it was the Hamas’ dream team . . . the ceiling collapsed, but the team got away.  No one knows how many Israelis were killed as a result of the decision [not to use heavier munitions].”[8]

III.  Choosing the Framework

A.  Justifying Targeted Killings—The War Paradigm

The debate within the United States over the lawfulness of targeted killings has remained largely confined to legal scholarship and public commentary; the courts have never addressed it.  The Bush administration, to a large extent, relied on a December 1989 Memorandum of Law (an advisory opinion), issued by the Special Assistant for Law of War Matters to The Judge Advocate General of the Army at the time, W. Hays Parks.[9] The Parks memorandum distinguished the prohibition on illegal assassinations in Executive Order 12333 from lawful targeting of individuals or groups who pose a direct threat to the United States.  The prohibition, argued Parks, applied to covert acts of murder for political reasons.  Legal Advisor to the State Department at the time, Arbaham Sofaer, emphasized in his own statements that the prohibition “should not be limited to the planned killing only of political officials, but that it should apply to the illegal killing of any person, even an ordinary citizen, so long as the act has a political purpose.”[10] Both Parks and Sofaer, however, asserted that this prohibition did not preclude the targeted killing of enemy combatants in wartime or the killing in self-defense of specific individuals who pose a direct threat to U.S. citizens or national security in peacetime.  The latter, both argued, was permissible under the inherent principle of self-defense to which every country was entitled under Article 51 of the United Nations Charter (which allows countries to use force in self-defense after suffering an “armed attack”) and customary international law.  Neither Parks nor Sofaer expounded on what amounts to direct threat.

The Bush administration has favored the paradigm of war, treating terrorists as combatants and justifying the targeted killing of terrorists as equivalent to the lawful killing of members of an enemy force on any battlefield.  Specifically, the administration deemed terrorists to be “unlawful combatants,” targetable and detainable, but denied the rights accorded to lawful detainees, namely, to be treated as prisoners of war if captured.  The Bush administration maintained this position even when the targeted killing took place in Yemen or Pakistan, outside an immediate theater of hostilities such as Afghanistan.  Given that the war on terrorism was a “global war,” the Administration maintained, there could be no geographical boundaries to the theater of war.

However, as we noted in the introduction to this chapter, choosing a war paradigm as governing the targeted killings of suspected terrorists is not devoid of difficulties.  The killing on the basis of blame rather than status, the difficulties in ensuring the accurate identification of the target, and the fact that operations take place outside of a defined battlefield—all make the war paradigm at best a proximate, but by no means a perfect, fit.  The full legal implications of this choice were considered by the Israeli High Court of Justice (HCJ), in its ruling on the Israeli practice of targeted killing operations in Gaza and the West Bank.

A petition was first submitted to the HCJ by a group of Israeli NGOs in late 2001, as the first Israeli targeted killing operations became public, but it was dismissed on grounds of justiciability.  In March 2002, another petition was submitted, and this time, Supreme Court President Aharon Barak ordered the state to respond.  By that time, 339 Palestinians had been killed by targeted killing operations during the Second Intifada: 201 intended targets and 129 innocent bystanders.  No less than seven briefs, covering hundreds of pages of arguments and documents, were submitted to the court.  In his last decision before retiring from the court, President Barak delivered the ruling in December 2006.[11] It is probably the most comprehensive judicial decision ever rendered addressing the legal framework of the “war on terrorism.”

Barak began by accepting that, unlike in the era of the First Intifada, there was now an “international armed conflict” with Palestinian militants, which warranted and justified the use of military means, as governed by customary international law, to combat terrorism.  For Barak, accepting the armed conflict paradigm was, albeit implicitly, a precondition to the justification of targeting operations, going far beyond any law enforcement method.  Furthermore, his choice of an international armed conflict paradigm was singular amongst the opinions of the U.S. Supreme Court as well as most other commentators, which have favored a non-international armed conflict model.  This choice was possibly motivated by the fact that international armed conflicts are subject to more regulation under international law than their non-international counterparts, thereby further constraining the government.

Barak, in his decision, did not discuss the possibility of working within a law enforcement paradigm, or the possibility of relying on Article 51 of the UN Charter to justify the practice.  Indeed, it would have been hard to justify a general practice, employed hundreds of times in the same territory, as an “exceptional measure” under a self-defense paradigm.

But Barak’s acceptance of the war paradigm as applicable to the fight against terrorism was not unqualified.  The remainder of the decision was designed to limit the full application of the laws of war and place further constraints on the legitimacy of targeted killing operations, in comparison with traditional combat.

First, in terms of the classification of terrorists, Barak rejected the government’s claim that these were unlawful combatants, and found, instead, that terrorists were “civilians taking direct part in hostilities.”  This choice of a two-group classification (civilian/combatant) vs. a three-group classification (civilian/ lawful combatant/unlawful combatant) was intended to achieve at least two goals.  The first was to make sure the protections of the Geneva Conventions applied to the armed conflict with Palestinian terrorists and to avoid the American administration’s conclusion that, as “unlawful combatants,” terrorists were entitled to few protections under the laws of war.

Second, by refraining from labeling terrorists as “combatants,” the ruling ensured that unlike combatants on the battlefield, who were all legitimate targets regardless of rank, role, or threat, terrorists would not be targeted on the basis of mere membership in a terrorist organization; instead, an individual culpability of the targeted person, by way of direct participation in instigating and executing terrorist acts, would have to be proven.  A mere membership test in the case of Hamas or some other Palestinian organizations would have been especially prone to over-inclusive application, as alongside their military wings, these organizations also have broad political, social, economic, and cultural operations.

The ruling also departed from the traditional armed conflict paradigm in that it conditioned the legitimacy of targeted killings on the absence of a reasonable alternative for capturing the terrorist.  On the traditional battlefield, no attempt to capture the enemy or warn the enemy in advance is necessary before shooting to kill.  In fact, the court’s requirement to try to apprehend the terrorist is far more easily situated within a law enforcement model of regular policing operations and signifies the uneasiness that the court felt about the war paradigm.

The Supreme Court’s decision also addressed concerns about collateral damage to innocent civilians in the course of targeted killings operations.  At the time the petition was submitted, the ratio of civilians to militants killed by targeted killings operations was 1:3—one civilian for every three militants[12] (the ratio has improved substantially since then, and in 2007, the rate of civilians hurt in targeted killing operations was 2–3 percent).[13] Barak acknowledged the difficulty in determining what number of casualties was “proportionate”:

[O]ne must proceed case by case, while narrowing the area of disagreement.  Take the usual case of a combatant, or a terrorist sniper shooting at soldiers or civilians from his porch.  Shooting at him is proportionate even if as a result, an innocent civilian neighbor or passerby is harmed.  This is not the case if the building is bombed from the air and scores of its residents and passerby are harmed. . . .  The hard cases are those which are in the space between the extreme examples.[14]

Accordingly, the decision placed an emphasis on the procedure by which the targeted killing operation was considered and approved and on the post-factum debriefing of operations, all in an effort to improve the record on collateral harm.  Importantly, however, the decision did not demand a zero civilian casualty policy.  In that, it remained more loyal to the war paradigm than to a policing paradigm.

Barak added that certain incidents might be subjected to judicial review.

The concern about collateral damage also brought the court to stipulate that in certain cases in which there was substantial collateral damage, and depending on the conclusions of an investigation into such incidents, it would be appropriate to compensate innocent civilians who have been harmed.[15]

To conclude, the Israeli Supreme Court sought a middle ground between a more aggressive law enforcement paradigm and a tamer wartime paradigm.  It chose the latter as its point of departure, but then, in consideration of the unique nature of the war on terrorism, added limitations and constraints on the government’s war powers so as to remain as loyal as possible to the basic principles and values of the Israeli legal system.[16]

B.  Justifying Targeted Killings—The Exceptional Peacetime Operations Paradigm

Could the U.S. administration, given the Parks memorandum, justify targeted killings even without relying on the applicability of military powers to a “war on terrorism”?  It would have to find the operation lawful under a reasonable interpretation of the domestic law of homicide; it would have to address major issues of peacetime international law, including human rights laws and the duty to respect the sovereignty of other countries; and, of course, it would have to satisfy the constitutional protections found in the Bill of Rights, to the extent these are applicable abroad.

Domestic law enforcement operations permit shooting to kill a suspected criminal only under very limited circumstances.  These limitations coincide with international human rights norms on the use of force by governments against citizens.  When the suspect imposes no immediate and lethal threat, firing at him to affect an arrest is only constitutional if “the officer has probable cause to believe that the suspect poses a threat of serious physical harm.”[17] There are even greater common law constraints on shooting a suspect where there is a concern about collateral harm to others around the suspect; in such cases, law enforcement officials are required to hold their fire and refrain from risking innocent bystanders.  Still, under the American Model Penal Code § 3.02, the defense of “necessity” or “choice of evils” justifies and thus immunizes conduct “which the actor believes to be necessary to avoid a harm or evil to himself or to another” if the harm to be avoided is greater than that sought to be prevented by the law defining the crime (intentional killing, in this case), and so long as there is no reason to believe the legislature intended to exclude this justification.  Under this statement of the American rule the danger of the harm to be avoided need not be imminent and the rule would justify homicide as well as less serious crimes.  Thus, in some jurisdictions the wording need hardly be stretched to make legal under domestic law the killing of an active participant in a terrorist scheme to kill many others, if that way of aborting the plan is believed to be necessary.  In other jurisdictions the law would have to be changed to allow intentional homicides or consideration of non-imminent harms.

As for international human rights laws, the possibility of using deadly force against individuals who are threatening the security of the state has not been rejected altogether even by international human rights bodies.  The Human Rights Committee, in its response to the Israeli report on the practice of targeted killings, notes only that “[b]efore resorting to the use of deadly force, all measures to arrest a person suspected of being in the process of committing acts of terror must be exhausted.”[18] It adds that such operations must never be carried out for purposes of retribution or revenge, thus implying that they may be legitimate if intended at preemption.

The 2002 Inter-American Commission on Human Rights Report on Terrorism and Human Rights[19] similarly leaves room for the use of deadly force against suspected terrorists, even under a general law enforcement model.  It notes that “in situations where a state’s population is threatened by violence, the state has the right and obligation to protect the population against such threats and in so doing may use lethal force in certain situations.”[20] It goes on to assert the natural implication that, in their law enforcement initiatives, “states must not use force against individuals who no longer present a threat as described above, such as individuals who have been apprehended by authorities, have surrendered or who are wounded and abstain from hostile acts.”[21]

And in its decision in the case of Isayeva, the European Court of Human Rights acknowledged the right of a state—Russia—to use deadly force against Chechen rebels, even when there was no indication that the latter were posing an immediate threat to the Russian forces.[22]

But outside the territory of the United States, the government is also limited by the international norms protecting each state’s sovereignty in using force to capture or kill suspected criminals.  As a general principle of international law, a country is strictly prohibited from engaging in law enforcement operations in the territory of another country, and much more so when the law enforcement operation includes killing a person.  Deadly attacks by air strikes or drones directly implicate the international prohibition on the use of force between states.  How, then, could the government justify targeted killing operations under international law in any way other than relying on a war/combatants paradigm?

The Parks memorandum addresses the question of lawful targeting and unlawful assassinations in peacetime, and argues the following:

The use of force in peacetime is limited by the previously cited article 2(4) of the Charter of the United Nations.  However, article 51 of the Charter recognizes the inherent right of self-defense of nations.  Historically, the United States has resorted to the use of military force in peacetime where another nation has failed to discharge its international responsibilities in protecting U.S. citizens from acts of violence originating in or launched from its sovereign territory, or has been culpable in aiding and abetting international criminal activities.[23]

Parks goes on to give the examples of an 1804–1805 Marine expedition into Libya to capture or kill the Barbary pirates; a year-long campaign in 1916 into Mexico to capture or kill the Mexican bandit Pancho Villa following Villa’s attack on Columbus, New Mexico; the 1928–1932 U.S. Marines’ campaign to capture or kill the Nicaraguan bandit leader Augusto Cesar Sandino; the Army’s assistance in 1967 to the Bolivian Army in its campaign to capture or kill Ernesto “Che” Guevara; the forcing down in 1985 of an Egypt Air plane in Sicily, in an attempt to prevent the escape of the Achille Lauro hijackers; and the 1986 attacks on terrorist-related targets in Libya.

These historical precedents, claims Parks, support the interpretation of the United Nations Charter as authorizing the use of military force to capture or kill individuals whose peacetime actions constitute a direct threat to U.S. citizens or national security.  In a footnote, he adds:

In the employment of military force, the phrase “capture or kill” carries the same meaning or connotation in peacetime as it does in wartime.  There is no obligation to capture rather than attack the enemy.  In some cases, it may be preferable to utilize ground forces to capture (e.g.) a known terrorist.  However, where the risk to U.S. forces is deemed too great . . . it would be legally permissible to employ (e.g.) an air strike against that individual or group rather than attempt his, her, or their capture, and would not constitute assassination.[24]

If so, targeted killings, as they have been used by the United States in Yemen, Pakistan, and elsewhere, may well have been justified without ever relying on a “war on terrorism,” but instead by being framed as an exceptional use of force in self-defense alongside peacetime law enforcement.  Although Parks does not expound upon this point, from his equation of military action in peacetime with that of wartime, it seems he would accept some degree of collateral damage in a peacetime operation under similar logic of a wartime attack.

Choosing a peacetime framework with some allowance for military action is not free from difficulties.  One obvious problem is that the “exceptional” use of force has been turned, in the context of the war on terrorism, into a continuous practice.  In addition, the degree to which countries should be allowed to use force extraterritorially against non-state elements has been debated extensively by both international law and domestic law scholars.  The implications of allowing the use of armed force to capture or kill enemies outside a country’s own territory, and outside a theater of traditional armed conflict, may include spiraling violence, the erosion of territorial sovereignty, and a weakening of international cooperation.

Once the precedent is laid for a broad interpretation of Article 51 of the UN Charter, as existing alongside or as an exception to normal peacetime limitations, it becomes harder to distinguish what is allowed in peace from what is allowed in war.  It is for these reasons that not everyone accepts Parks’ legal reasoning, with critics arguing that any military attack on another country’s territory, outside an armed conflict with that country, amounts to unlawful aggression.  Thus, in the case of Armed Activities on the Territory of the Congo[25], the International Court of Justice, in a decision widely criticized, went as far as to rule that Uganda had no right to use force against armed rebels attacking it from the territory of the Democratic Republic of Congo.  Recently, the U.N. Special Rapporteur on Extrajudicial, Summary, or Arbitrary Executions, concluded that reliance on the exceptional self-defense argument under Article 51 in support of targeted killings “would diminish hugely the value of the foundational prohibition contained in Article 51.”[26]

Even if justified as an exception to a peacetime paradigm, one obvious precondition for the legality of targeted killing operations outside a theater of war, in consideration of the other countries’ sovereignty, must be that the state in whose territory the terrorist resides either consent to the operation by the foreign power (as in the case of the collaboration between the United States and Yemen) or else would be unable or unwilling to take action against the terrorist (as in the case of targeted killings in Gaza).  On some rare occasions there may be an overwhelming necessity to act without the immediate possibility of obtaining the other country’s consent.

Note that under a law enforcement model, a country cannot target any individual in its own territory unless there is no other way to avert a great danger.  If so, if the Yemeni authorities can capture a terrorist alive, they cannot authorize the United States to engage in a targeted killing operation in its territory or execute one on its own.

To sum up, targeted killings of terrorists by both the United States and Israel have been justified under a war paradigm: in the American case, by treating terrorists as (unlawful) combatants; in the Israeli case, by treating terrorists as civilians who are taking direct part in hostilities.  It seems that a persuasive argument can also be made that under some conditions, targeted killings of suspected terrorists can be justified on the basis of a law enforcement paradigm.  When conducted in the territory of another country, targeted killing must be based on a self-defense exception to the international law prohibition on the use of force, and in consideration of that other country’s sovereignty, should only be executed if that other country either consents to the operation or else is unable and unwilling to interdict the terrorist.

In the conclusion of this chapter, we set forth what the legitimate contours of the use of targeted killing must be.  We conclude that they seem to fit both a more constrained war paradigm and a more lax law enforcement paradigm (although the latter suits more sporadic and measured use of the tactic).  For present purposes it should be noted that if we take the Israeli Supreme Court’s decision as controlling, then the conditions for the legitimacy of targeted killings of terrorists in the armed conflict between Israel and Palestinian militants are not very different from those that would apply under a law enforcement model.  Both would allow the targeted killing of some terrorists in Gaza and both would prohibit—or place greater constraints—on the targeting of suspected terrorists outside a conventional theater of war if the alternative of capture was feasible.

IV.  Strategic Aspects

Even if legally justifiable and morally permissible, the strategic value of employing targeted killings is far from clear and depends very much on the situation.  As with any other counterterrorism tactic, targeted killings carry both strategic benefits and costs.

A.  The Potential Hazards of Targeted Killings

An immediate consequence of eliminating leaders of terrorist organizations will sometimes be what may be called the Hydra effect, the rise of more—and more resolute—leaders to replace them.  The decapitating of the organization may also invite retaliation by the other members and followers of the organization.  Thus, when Israel assassinated Abbas Mussawi, Hezbollah‘s leader in Lebanon, in 1992, a more charismatic and successful leader, Hassan Nassrallah, succeeded Mussawi.  The armed group then avenged the assassination of its former leader in two separate attacks, blowing up Israeli and Jewish targets in Buenos Aires, killing over a hundred people and injuring hundreds more.

Targeted killing may also interfere with important gathering of critical intelligence.  The threat of being targeted will drive current leaders into hiding, making the monitoring of their movements and activities by the counterterrorist forces more difficult.  Moreover, if these leaders are found and killed, instead of captured, the counterterrorism forces lose the ability to interrogate them to obtain potentially valuable information about plans, capabilities, or organizational structure.

The political message flowing from the use of targeted killings may be harmful to the attacking country’s interest, as it emphasizes the disparity in power between the parties and reinforces popular support for the terrorists, who are seen as a David fighting Goliath.  Moreover, by resorting to military force rather than to law enforcement, targeted killings might strengthen the sense of legitimacy of terrorist operations, which are sometimes viewed as the only viable option for the weak to fight against a powerful empire.  If collateral damage to civilians accompanies targeted killings, this, too, may bolster support for what seems like the just cause of the terrorists, at the same time as it weakens domestic support for fighting the terrorists.

When targeted killing operations are conducted on foreign territory, they run the risk of heightening international tensions between the targeting government and the government in whose territory the operation is conducted.  Israel’s relations with Jordan became dangerously strained following the failed attempt in September 1997 in Jordan to assassinate Khaled Mashaal, the leader of Hamas.  Indeed, international relations may suffer even where the local government acquiesces in the operation, but the operation fails or harms innocent civilians, bringing the local government under political attack from domestic constituencies (recall the failed attack in Pakistan on Al-Zawahiri that left eighteen civilians dead).

Even if there is no collateral damage, targeted killings in another country’s territory threatens to draw criticism from local domestic constituencies against the government, which either acquiesced or was too weak to stop the operation in its territory.  Such is the case now in both Pakistan and Yemen, where opposition forces criticize the governments for permitting American armed intervention in their countries.

The aggression of targeted killings also runs the risk of spiraling hatred and violence, numbing both sides to the effects of killing and thus continuing the cycle of violence.  Each attack invites revenge, each revenge invites further retaliation.  Innocent civilians suffer whether they are the intended target of attack or its unintentional collateral consequences.

Last but not least, exceptional measures tend to exceed their logic. As in the case of extraordinary detention or interrogation methods, there is a danger of over-using targeted killings, both within and outside of the war on terrorism.  A particular danger in this context arises as the killing of a terrorist often proves a simpler operation than protracted legal battles over detention, trial, extradition, and release.

B.  The Benefits Nations Seek

At the most basic level, targeted killings, which are generally undertaken with less risk to the attacking force than are arrest operations, may be effective.  According to some reports, the killing of leaders of Palestinian armed groups weakened the will and ability of these groups to execute suicide attacks against Israelis.  By deterring the leaders of terrorist organizations and creating in some cases a structural vacuum, waves of targeted killing operations were followed by a lull in subsequent terrorist attacks, and in some instances, brought the leaders of Palestinian factions to call for a ceasefire.  The Obama administration embraced the targeted killing tactic, holding it to be the most effective way to get at Al-Qaeda and Taliban members in the ungoverned and ungovernable tribal areas along the Afghanistan-Pakistan border or in third countries.

Despite the adverse effects such operations may have on the attitudes of the local population toward the country employing targeted killings, the demonstration of superiority in force and resolve may also dishearten the supporters of terrorism.

Publicly acknowledged targeted killings are furthermore an effective way of appeasing domestic audiences, who expect the government “to do something” when they are attacked by terrorists.  The visibility and open aggression of the operation delivers a clearer message of “cracking down on terrorism” than covert or preventive measures that do not yield immediate demonstrable results.  The result in Israel has been to make a vast majority of citizens supportive of targeted killings, despite the latter’s potential adverse effects.  And, perhaps surprisingly, of all the coercive counterterrorism techniques employed by the United States, targeted killings have so far attracted the least public criticism.

V.  Conclusions

Targeted killing operations display more clearly than any other counterterrorism tactic the tension between labeling terrorism a crime and labeling it an act of war.  If a terror attack is simply a crime, counterterrorism forces would follow the same laws and rules as the Chicago or Miami police department do in fighting crime, where intentional killing could rarely if ever be lawful, other than where necessary in a situation immediately requiring the defense of self or others, or in making an arrest of an obviously dangerous felon.  From the perspective of international peacetime relations, targeted killings face even greater legal constraints when targeting a terrorist outside the state’s jurisdiction.

If a terrorist plan is an act of war by the organization supporting it, any member of any such terrorist organization may be targeted anytime and anywhere plausibly considered “a battlefield,” without prior warning or attempt to capture.

Known or anticipated collateral damage to the innocent is generally prohibited in law enforcement, but is legitimate within the boundaries of proportionality in fighting wars.  In fighting crime, the government’s obligation to protect its citizens applies to all citizens—criminals and innocents.  In fighting wars, the government’s primary obligation is to its own citizens, with only limited concern for the well-being of its enemies.

Assuming, as we do, that states do have a right to defend themselves against acts of terrorism, targeted killings cannot be always illegal and immoral.  But because terrorism is not a traditional war, nor a traditional crime, its non-traditional nature must affect the ethical and strategic considerations that inform targeted killings, the legal justification behind them, and the choice of targets and methods used to carry them out.

As we have shown, targeted killings may be justified even without declaring an all-out “war” on terrorism.  A war paradigm is overbroad in the sense that it allows the targeting of any member of a terrorist organization.  For the United States, it has had no geographical limits.  When any suspected member of a hostile terrorist organization—regardless of function, role, or degree of contribution to the terrorist effort—might be targeted anywhere around the world without any due process guarantees or monitoring procedures, targeted killings run grave risks of doing both short-term and lasting harm.  In contrast, a peacetime paradigm that enumerates specific exceptions for the use of force in self-defense is more legitimate, more narrowly tailored to the situation, offers potentially greater guarantees for the rule of law.  It is, however, harder to justify targeted killing operations under a law enforcement paradigm when the tactic is used as a continuous and systematic practice rather than as an exceptional measure.  Justifying targeted killings under a law enforcement paradigm also threatens to erode the international rules that govern peacetime international relations as well as the human rights guarantees that governments owe their own citizens.

Whichever paradigm we choose as out starting point, greater limitations than those offered by the Parks memorandum or that are currently operating in the American targeted killings program should be adopted.  The limits set by the Israeli Supreme Court—ironically, within the paradigm of wartime operations—are a good place to start.

First, the tactic should not be used unilaterally by the endangered state if the host country of the terrorists is willing and able to act on its own to arrest or disable in a timely manner the source of the threat.  Host country cooperation in capture and extradition must be the first alternative considered.  That is, targeted killings must only be carried out as an extraordinary measure, where the alternative of capture or arrest is unfeasible.

Second, only those who are actively and directly involved in terrorist activities are legitimate targets; not every member of a terrorist organization is or should be.

Third, the fact that terrorists do not wear uniforms should not give them an unfair legal advantage over soldiers in uniform in the sense of immunity from deliberate attack.  But their lack of uniform does raise legitimate concerns about the ability to ensure the correct identification of the target, in terms of personal identity as well as specific culpability.  Any targeted killing operation must therefore include mechanisms in its planning and execution phases that would ensure an accurate identification.  Such mechanisms need not involve external judicial review; judges are neither well situated nor do they have the requisite expertise to authorize or reject an operation on the basis of intelligence reports.  Rather, the system should be based on verified and verifiable intelligence data from different and independent sources, careful monitoring, and safety mechanisms that would allow aborting the mission in case of doubt.

The concern about collateral damage requires specific attention.  Unlike ordinary battlefield strikes, the fact that the targeting forces have control over the time, means, and methods of strike mandates that a heightened degree of care should be exercised to choose an occasion and means that will minimize collateral harm to uninvolved individuals, especially where the operations are carried out outside an immediate conflict zone.  In those cases, we believe that where innocent civilians suffer collateral damage, those injured should generally be compensated.

Finally, the aggression of the targeted killing tactic mandates its measured use in only the most urgent and necessary of cases.  The government’s interest should be to tame violence, not exacerbate it.  Where alternatives exist, they should be pursued, not just as a matter of law but also as a matter of sound policy.

This article appears as a chapter in Gabriella Blum & Philip Heymann, Laws, Outlaws, and Terrorists: Lessons from the War on Terrorism (MIT Press, forthcoming Sept. 2010).


*Assistant Professor of Law, Harvard Law School.

** James Barr Ames Professor of Law, Harvard Law School.

[1] One such famous case took place in July 1973, when the Israeli Mossad assassinated an innocent Moroccan waiter in Lillehammer, Norway, mistaking him for a member of the Black September faction responsible for the Munich massacre.

[2] For a report on the U.S. “hit list,” see Craig Whitlock, Afghans Oppose U.S. Hit List of Drug Traffickers, Wash. Post, Oct. 24, 2009, available at http://www.washingtonpost.com/wp-dyn/content/article/2009/10/23/AR2009102303709.html.

[3] Dana Priest, U.S. Military Teams, Intelligence Deeply Involved in Aiding Yemen on Strikes, Wash. Post, Jan. 27, 2010, available at http://www.washingtonpost.com/wp-dyn/content/article/2010/01/26/AR2010012604239.html.

[4] Id.

[5] The latter flew to London in September 2005 following his discharge from the military, but had to stay aboard the plane and return to Israel after being tipped off that he might be arrested.

[6] Matar v. Dichter, 500 F. Supp. 2d 284 (S.D.N.Y. 2007).

[7] The United States Court of Appeals for the Second Circuit affirmed the dismissal. See Matar v. Dichter, 563 F.3d 9 (2d Cir. 2009).

[8] Amos Harel, Dichter: The Targeted Killing of Hamas Leaders Has Brought About Calm, Ha’aretz, June 1, 2005 (in Hebrew).

[9] Memorandum from W. Hays Parks, Special Assistant to The Judge Advocate Gen. of the Army for Law of War Matters, to The Judge Advocate Gen. of the Army, Executive Order 12333 and Assassination (Dec. 4, 1989) reprinted in Army Law., Dec. 1989, at 4 [hereinafter Parks Memorandum] available at http://www.loc.gov/rr/frd/Military_Law/pdf/12-1989.pdf.

[10] Abraham D. Sofaer, Sixth Annual Waldemar A. Solf Lecture in International Law: Terrorism, the Law, and the National Defense, 126 Mil. L. Rev. 89, 119 (1989).  For further analysis of the Sofaer doctrine, see Kenneth Anderson, Targeted Killing in U.S. Counterterrorism Strategy and Law 24–25 (Brookings Inst., Georgetown Univ. Law Ctr., & Hoover Inst., Working Paper of the Series on Counterterrorism and American Statutory Law, 2009), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1415070.

[11] HCJ 769/02 Public Comm. Against Torture in Israel v. Gov’t of Israel (Targeted Killings Case) [2005], available at http://elyon1.court.gov.il/Files_ENG/02/690/007/a34/02007690.a34.pdf.

[12] Note that numbers of militants killed include both the intended targets and their armed group associates who were present at the time of the attack and were harmed as a result.

[13] Amos Harel, Pinpointed IAF Attacks in Gaza More Precise, Hurt Fewer Civilians, Ha’aretz, Dec. 30, 2007, available at http://www.haaretz.co.il/hasen/spages/939702.html.

[14] Targeted Killings Case, HCJ 769/02 at ¶ 46.

[15] Id. at ¶ 40.

[16] That the 2008 armed conflict between Israel and Hamas in Gaza looked far more like a conventional war may help explain why, only three years after Barak’s decision, Israeli forces struck numerous Hamas members who would not have necessarily met the strict tests he had imposed.

[17] Tennessee v. Garner, 471 U.S. 1 (1985).

[18] U.N. Human Rights Committee, Concluding Observations of the Human Rights Committee: Israel, ¶ 15, U.N. Doc. CCPR/CO/78/ISR (Aug. 21, 2003), available at http://www.unhchr.ch/tbs/doc.nsf/%28Symbol%29/CCPR.CO.78.ISR.En?OpenDocument.

[19] Inter-American Commission on Human Rights, Report on Terrorism and Human Rights, OEA/Ser.L/V/II.116, Doc. 5 Rev. 1 Corr (Oct. 22, 2002), available at http://www.cidh.oas.org/Terrorism/Eng/exe.htm.

[20] Id. ¶ 87.

[21] Id. ¶ 91.

[22] Isayeva v. Russia, 41 Eur. Ct. H. R. 847 ¶ 181 (2005); see also Isayeva v. Russia, App. No. 6846/02, Eur. Ct. H. R. (Nov. 15, 2007).

[23] Parks Memorandum, supra note 9, at 7.

[24] Id. at 8 n.14.

[25] Armed Activities on the Territory of the Congo (Dem. Rep. Congo v. Uganda), 2005 I.C.J. 116 (Dec. 19).

[26] Special Rapporteur on Extrajudicial, Summary or Arbitrary Executions, Report of the Special Rapporteur on Extrajudicial, Summary or Arbitrary Execution: Study on Targeted Killings, ¶ 41, delivered to the Human Rights Council, U.N. Doc. A/HRC/14/24/Add.6 (May 28, 2010), available at http://www2.ohchr.org/english/bodies/hrcouncil/docs/14session/A.HRC.14.24.Add6.pdf.

Assistant Professor of Law, Harvard Law School.

James Barr Ames Professor of Law, Harvard Law School.