Lethal Force at Home and Abroad


Daniel B. Loehr


This Article compares the legal treatment of lethal force exercised by the President with the legal treatment of lethal force exercised by police officers. In doing so, it examines the distinction between the crime paradigm and the war paradigm, complicates assumptions about when our rights as citizens are the strongest, and offers lessons about how to create meaningful constraints for lethal force. The Article concludes descriptively that when contemplating lethal force, the President is subject to more constraints than are police officers. Two insights follow from this claim. First, lethal force is an exception to the assumed rule of broader rights and limited government in the crime paradigm. Second, the relative utility of ex ante regulation for the President provides yet another reason to improve ex ante regulation for the police.

I. Introduction

The President authorizes a drone strike against a suspected terrorist outside of active hostilities. A police officer shoots a man who allegedly threatens him. How do the respective legal treatments differ? While lethal force by the President and lethal force by the police pose similar regulatory challenges,1 the literature so far has analyzed them in isolation.2 In comparing the two, this Article observes that the tools of criminal and constitutional law do not constrain the police more than the President. Moreover, because the President is constrained by alternative means—such as international law, authorizing statutes, and internal policy—the net result is more constraint for the President. Understanding this yields lessons about the nature of our rights in different contexts and suggests that regulatory tools traditionally used to constrain the otherwise unbound President may be useful for regulating police lethal force as well.

Consider an example from each context. On September 30, 2011, the CIA killed Anwar al-Awlaki, a U.S. citizen born in New Mexico, by directing Hellfire missiles at his car in northern Yemen.3 Announcing the strike later that day, President Obama described Awlaki as “the leader of external operations for al-Qaeda in the Arabian Peninsula.”4 In that role, Awlaki is alleged to have directed the failed attempt to blow up a passenger plane in the United States in 2009 and the failed attempt to blow up U.S. cargo planes in 2010.5 In killing Awlaki, the U.S. government asserted that the threat he posed was imminent and that capturing him was infeasible.6

Nearly three years later on August 9, 2014, a Ferguson, Missouri police officer killed Michael Brown, a U.S. citizen born in Missouri,7 by shooting him six to eight times with a Sig Sauer pistol.8 Discussing the shooting a day later, St. Louis County Police Chief Jon Belmar described Brown as having “assaulted” a police officer.9 The officer who shot Michael Brown later alleged that the shooting was justified because Brown posed an imminent threat to his life, and that he could not safely incapacitate Brown by other means.10 Despite the factual similarities of these killings, their respective legal treatments diverge dramatically.11 The two uses of force are placed in largely separate regulatory worlds, falling on opposite sides of a controlling conceptual line—that dividing the crime paradigm and the war paradigm.12 While these modes of lethal force do not represent the archetypes of the dichotomy between the paradigms (as might the death penalty and the killing of citizens in war), they do fall distinctly into the two categories. In regulatory, popular, and academic contexts, police lethal force is situated in the sphere of domestic law enforcement and executive lethal force is situated in the sphere of international military affairs. In many areas of law, the crime/war distinction determines the extent to which government action is regulated. Citizens are generally more protected, and the government more restricted, in a domestic crime context than in an international military context. For example, consider the requirements of detention. In the crime context, prolonged detention of a citizen inheres a right to trial and requires a determination of guilt.13 In the war context, however, the burden is placed on the citizen-detainee to challenge her detention.14 The same goes for evidentiary requirements at trial, which are relaxed for a citizen captured on a battlefield abroad.15 This dynamic can be justified on numerous grounds, including capacity (e.g. that the tools of constitutional and criminal enforcement afford greater rights protection domestically) and necessity (e.g. that we give broader discretion to leaders in the war context because they ostensibly protect against greater collective harms). When cast on lethal force, the crime/war distinction predicts that restrictions on the police would be greater than restrictions on the President.

However, this relationship is inverted in the context of lethal force. When evaluating whether a particular exercise of lethal force will be lawful, the President faces greater hurdles than does the average police officer. This is so in at least four ways: Unlike the police, the constraints on the Executive require that force only be used as a last resort, that only the minimum required force is used, and that the effects of lethal force on bystanders are considered.16 Additionally, while both standards have an imminence requirement17 in name, the executive imminence requirement is stricter.18 Therefore, citizens have more protection if lethal force is contemplated by the President instead of the police.

This is a simplified account of the differences. The remainder of the Article makes the comparison at a more granular level. It proceeds by considering how various regulatory tools (criminal law, constitutional law, international law, authorizing statutes, and internal policy) operate on the police and the President. Part II contends that constitutional law and criminal law provide a similarly minimal amount of constraint on both the President and the police. Looking beyond constitutional law and criminal law, Part III identifies where the regulatory systems diverge: While police regulation largely rests on ineffective constitutional and criminal law, regulation of the President is supplemented by an additional infrastructure of constraint—namely international law,19 authorizing statutes,20 and internal policy.21 It is this set of supplemental regulatory tools that accounts for the more robust constraint on the President; Part IV considers how to translate them to the police context. On the whole, the analysis indicates that while the President and the police are similarly unconstrained by constitutional and criminal law, the President is more significantly constrained by alternative means.

Recognizing this yields a number of benefits. First, it complicates assumptions about where and when rights are strongest. The example of lethal force puts pressure on the general account that rights are more robust and the government is more constrained in the crime paradigm and counsels against arguments that presuppose the superiority of crime paradigm regulatory tools without further functional analysis. Constitutional and criminal law should be forced to earn their reputation as superior regulatory tools.

Second, by viewing police lethal force in light of lethal force by the President, this Article draws on a heretofore untapped resource to contribute to the debate about how to regulate police lethal force.22 Put differently, what does regulation of the President teach about regulating the police? The President has become more constrained than the police because she is not just limited by constitutional and criminal law, but also by norms of international law,23 ex ante authorizing statutes,24 and internal policy.25 It is this type of ex ante regulation that is missing in the police context. In so arguing, this Article lends support to the body of literature calling for ex ante regulation of the police,26 and further suggests that to the extent ex ante regulation is needed for policing generally, it is especially so for lethal force.

At bottom, the suggestion of this Article is modest. It argues descriptively that when contemplating lethal force, the Executive is subject to constraints that police officers are not. Two insights follow from this claim. First, lethal force is an exception to the assumed rule of broader rights and limited government in the crime paradigm. Second, the relative utility of ex ante regulation for the President provides yet another reason to consider using ex ante regulation for the police.

II. Underperforming Crime Paradigm Regulatory Tools

It is generally believed (correctly) that criminal law and constitutional law provide only weak constraints on the President’s use of lethal force.27 It is also believed (incorrectly) that they provide stronger constraints on the police.28 This is because police have more formal risks of liability. When police use lethal force, they are theoretically subject to federal criminal liability, state criminal liability, and constitutional civil liability. By contrast, the President can reasonably assume that she will not face criminal or civil liability at any level.29 Criminal laws and constitutional laws of course still operate on the President, but functionally only through the Justice Department’s Office of Legal Counsel (OLC) as an ex ante check.30 Despite the opportunity afforded by constitutional and criminal enforcement, however, effective constraints have failed to materialize for the police. Therefore, contrary to conventional wisdom, criminal law and constitutional law provide only weak limits for both the police and the President.

A. Constitutional Law

While the Fourth Amendment applies to both the President and the police, the mode of application differs. In the executive context, courts decline to decide whether executive actions like the killing of Awlaki violate the Fourth Amendment,31 leaving the enforcement work to OLC.32 By contrast, courts at all levels apply the Fourth Amendment to the use of lethal force by police.33 This section considers whether constitutional law is more constraining on the police than on the President, and concludes that it is not.

1. The Police

The constitutional constraints imposed on the police use of force are weak. The Constitution prohibits the use of force only if it is “unreasonable.”34 This prohibition comes from the Fourth Amendment—protecting the right to be secure against unreasonable seizures.35 Because the Court has largely declined to specify the meaning of “reasonable,” the constitutional standard limiting the use of force remains indeterminate. In a 1985 case, Tennessee v. Garner, the Court came the closest to establishing clear rules for lethal force. Garner established the relatively clear rule that a police officer may not use lethal force against a fleeing felon unless “it is necessary to prevent the escape and the officer has probable cause to believe that the suspect poses a significant threat of death or serious physical injury to the officer or others.”36 Garner also required that warning be given “where feasible.”37 Together, these two rules represent the apex of clear constitutional limits on lethal force. But even these rules are neither entirely clear nor strict.

A weakness of Garner is that the Court did not specify what suffices for an officer to infer a “threat of . . . serious physical injury.”38 The Court did provide an example, but it creates rather than resolves ambiguity. “If the suspect threatens the officer with a weapon or there is probable cause to believe that he has committed a crime involving the infliction or threatened infliction of serious physical harm,” Garner held, “deadly force may be used if necessary to prevent escape.”39 This leaves open the possibility that an officer may use the fact of an already-completed violent crime to infer imminent danger. The implication is that if Edward Garner had not just broken into a home and stolen a purse, but also assaulted the resident of that home, the officer might have been justified in shooting him as he climbed the fence, despite the same facts that he was “young, slight, and unarmed.”40 Nonetheless, while this ambiguity weakens the protective force of Garner, it does not dissolve Garner’s central conclusion that it is constitutionally unreasonable to use lethal force against non-dangerous fleeing felons. Unfortunately, the Court’s path after Garner navigated further from clarity. In 1989, in Graham v. Connor,41 the Court crafted a standard for all use of force cases, not just the lethal force against a fleeing felon that was addressed in Garner. The standard set forth in Graham was objective reasonableness, to be determined by the totality of circumstances.42 To aid lower courts in the analysis, the Court listed some relevant factors to consider. In no particular order and with no specified weight, the factors were: “the severity of the crime at issue, whether the suspect poses an immediate threat to the safety of the officers or others, and whether he is actively resisting arrest or attempting to evade arrest by flight.”43

Because of this balancing, Graham represents a departure from anything rule-like about Garner. While Garner used balancing to produce a rule for lower courts, Graham simply instructed lower courts to balance. This undermines some of the rules set forth in Garner. For example, while Garner suggested that danger was a minimum requirement, Graham reduced danger to a non-dispositive factor to be balanced.44 The indeterminacy of Graham, especially with respect to imminent danger, poses a problem given that these cases double as instructional tools for officers.45 In addition, as discussed in Part II.C., the functional cost of indeterminacy increases exponentially in the context of qualified immunity. In Scott v. Harris, the Court moved even further from rules and into totality of circumstances balancing.46 In 2007, a Georgia police officer rammed the back of Victor Harris’ car to end a high-speed chase, leaving Harris a quadriplegic. Using a totality of circumstances balancing test, Scott held that the officer did not violate the Fourth Amendment and rejected the notion that the Fourth Amendment imposes specific requirements on the use of lethal force.47 In response to the plaintiff’s argument that the requirements of Garner were not met, the Court noted that, “Garner did not establish a magical on/off switch that triggers rigid preconditions whenever an officer’s actions constitute ‘deadly force.’ Garner was simply an application of the Fourth Amendment’s ‘reasonableness’ test.”48 Scott departed from Graham as well, because while Graham provided relevant factors to consider in the balancing, Scott simply instructed courts to “slosh [their] way through the factbound morass of ‘reasonableness.’”49 This is where the law stands today.

2. The President

The same Fourth Amendment standard applies to lethal force executed by the President. The only public application of the standard, however, comes with OLC’s analysis of the Awlaki strike, which found that the Fourth Amendment would not be violated by his killing.50 The memo drew reference to Garner,51 but rested its analysis on Scott’s totality of circumstances balancing. In balancing the totality of circumstances, the memo considered that the citizen was a member of an enemy force and was engaged in the continuous planning of attacks against the United States, that capture was unfeasible,52 and that high-level officials had made the decision (even though that is doctrinally irrelevant).53 The memo found that the balance weighed in favor of the government—perhaps not a shocking outcome given that on one side of the scale is the asserted prevention of mass violence against U.S. citizens. The OLC memo thus shows that like police officers using lethal force, all that the Fourth Amendment requires of the President is reasonableness, determined by the totality of circumstances.

Unlike the police, however, the President is also constrained by the Fifth Amendment; the OLC memo applied the Fifth Amendment to executive lethal force, even though it is not applied to the police.54 In applying the Fifth Amendment, the memo cited to Hamdi and applied the Mathews balancing analysis.55 Weighed broadly, the government interest was national security. More specifically, the memo noted that the suspect posed a “continued and imminent threat of violence or death” to U.S. persons.56 The citizen’s interest in the risk of an erroneous deprivation of his life, the memo noted, was “very real.”57 As for the possibility and burden of additional process, the memo noted that in the current process, “the highest officers in the intelligence community” reviewed the factual basis for the operation.58 The memo did not consider what additional process might look like, nor whether that process would have burdened the government. It did note that capture (and thus other procedures like a trial) would be unfeasible. In conclusion, the memo found that the weight of the government interest was such that the Constitution did not require additional process.59 Because the memo failed to consider the burden of additional process, it is difficult to view the application of the Fifth Amendment here as meaningfully constraining. Instead, the Fifth Amendment analysis accepted the weight of the government interest as pushing the required process to whatever the government sees fit, which in this case was high-level review. Since high-level review was found to be sufficient but not necessary, we do not know exactly what the Fifth Amendment demands for lethal force process. The extent to which the Fifth Amendment provides a constraint for the President that is not present for the police is therefore unclear. In sum, the constitutional protections against lethal force are weak for both the police and the President.

B. Criminal Law

1. The President

Despite the presence of applicable criminal laws, the President maintains significant leeway to use lethal force without committing a crime. The OLC memo that analyzed the legality of the Awlaki killing raised the possibility that three criminal statutes constrain executive use of lethal force.60 Ultimately, though, the memo found that none of them prohibited the Awlaki strike.61

The first such statute makes it illegal for a citizen to kill another citizen abroad.62 The public authority justification, however, exempts public officials when they act lawfully.63 “Lawfully” turns on analysis of authority and limits from constitutional and international law. In this way, while adherence to this statute requires constitutional and international law compliance, it does not impose independent limits for public officials. The same public official exemption applies to the second relevant statute, which prohibits conspiring within the United States to commit murder abroad.64 The third relevant statute, the War Crimes Act,65 does provide an independent outer limit for the operation. Pursuant to the War Crimes Act, the President is prohibited from using lethal force against someone who takes no active role in the hostilities, whether she is a prisoner of war or a former soldier who has now laid down arms.66 The OLC memo found, however, that this limitation did not apply to Awlaki, who was determined to be taking an active role in hostilities.67 Given the factual leeway to assert that somebody is taking part in hostilities, this criminal limitation on executive lethal force cannot be regarded as particularly strict. In sum, criminal laws give the Executive substantial leeway when it comes to the use of lethal force.

2. The Police

The role of criminal law is also minimal in the context of lethal force by the police. The federal vehicle for police criminal liability is 18 U.S.C. § 242, which enables federal prosecutors to charge law enforcement with willfully depriving a person of her constitutional rights while acting under color of law.68 Structured this way, an element of the criminal violation is a constitutional violation. Therefore, because prosecutors can only bring criminal cases when constitutional rights have been violated, federal criminal laws do not add to pre-existing constitutional limits. Neither do state criminal laws. While the constitutional standard does serve as the basis for civil liability and possible federal criminal liability, it does not control state criminal law.69 A state may both criminalize behavior that the Court deems constitutional and decline to criminalize behavior that the Court deems unconstitutional.70 Presently, state statutes fall into two categories: eleven states follow the common law rule71 and thirty-eight states follow the Garner rule.72 The common law rule permits the use of lethal force to make an arrest of a fleeing felon, making no distinction among felonies and not requiring dangerousness. The Garner rule, described above, requires that the felony be dangerous or that the felon be presently dangerous.73 Therefore, this tally demonstrates that as compared to the constitutional requirements for the use of lethal force, eleven states have less restrictive criminal law requirements and thirty-eight states have similarly restrictive criminal law requirements.74 In sum, for both the police and the President, criminal law does not impose meaningful limitations beyond the constitutional floor.

Therefore, constitutional and criminal law provide only weak constraints on both the President and the police. This is unsurprising in the executive context, but notable in the police context because criminal and constitutional liability are tools that lawyers and the public expect to meaningfully constrain police officers. Evidently, such an expectation is unwarranted.

C. Enforcing Constitutional and Criminal Law

The foregoing analysis points out that the criminal and constitutional lethal force standards are similarly weak for the police and the President. But of course, it is possible that the same weak standards are enforced more robustly against the police, thereby creating more constraint. In both the criminal and the constitutional context, this turns out not to be the case. In the constitutional context, the difference between enforcement by OLC (for the President) and enforcement by federal and state courts (for police officers) suggests that the probability of enforcement is higher in the police context. After all, courts have contempt power and OLC does not.75

But the reality is more complicated. The doctrine of qualified immunity diminishes the capacity for the Constitution to be enforced against police officers. Under qualified immunity law, an officer cannot be held liable for violating an individual’s rights unless, at the time of the incident, “it would [have been] clear to a reasonable officer that his conduct was unlawful.”76 Because the Court’s doctrine on lethal force remains non-specific, it is unlikely that an officer would find it “clear” that her behavior violates the Fourth Amendment. The Court admits as much. Brosseau v. Haugen concluded that Graham is not clear enough to provide officers with notice that would satisfy qualified immunity requirements.77 The interaction of the indeterminate constitutional standard and qualified immunity means that constitutional violations go unenforced. As Professor Harmon notes, indeterminacy is always unfortunate, but “in the context of 1983 litigation, because of qualified immunity, it is devastating.”78 Whether qualified immunity renders enforcement against the police as weak as OLC enforcement against the President cannot be said with certainty, but it undoubtedly weakens any enforcement advantages that § 1983 would otherwise afford.

Still, though, the nature of enforcement in court could provide other types of constraints on police officers not experienced by the President. Consider, for example, if lower courts particularized79 the general standard articulated by the Supreme Court. While OLC only applies the law articulated by the Supreme Court, if lower courts developed the general constitutional standard into more particular rules, police officers would be subject to them. But particularization has not come to pass. While examples of particularization do exist, the Court’s reasonableness standard has, on the whole, not been meaningfully developed. In Smith v. City of Hemet, for example, the Ninth Circuit introduced an additional factor to determine reasonableness: “the availability of alternative methods of capturing or subduing a suspect.”80 In Sharrar v. Felsing, the Third Circuit listed a number of possible factors to augment the reasonableness inquiry, including “the possibility that the persons subject to the police action are themselves violent or dangerous, the duration of the action, whether the action takes place in the context of effecting an arrest, the possibility that the suspect may be armed, and the number of persons with whom the police officers must contend at one time.”81 The Sixth Circuit took up the critical question of whether there are any minimum requirements for the lawful use of force. Reading Graham literally, there are none. But Sample v. Bailey established Graham’s threat factor as “a minimum requirement for the use of deadly force.”82 In another effort at particularization, the Ninth Circuit found that pre-seizure conduct was relevant to reasonableness.83 Last term, however, the Supreme Court rejected this effort, requiring a return to general reasonableness.84

Notwithstanding these few efforts at particularization, the general Graham/Scott practice of “sloshing” through the “morass” reigns.85 In large part, lower courts state the reasonableness test in Graham, note the relevant factors to be considered, and then apply (or occasionally ignore)86 the factors in their analysis.87 Thus, despite its potential, particularization has not brought about additional constraint for police officers.

Taken together, the strength of qualified immunity doctrine and the lack of particularization serve to mute any enforcement advantages that courts have over OLC enforcement. This is not to say that the differences in the mode of enforcement do not matter—they do. But the story is not one of robust enforcement in the police context and weak OLC enforcement in the executive context. It is instead a story of two imperfect enforcement regimes, applying the same, weak constitutional standard. In form and function, the Constitution fails to provide meaningful regulation for police and executive lethal force. While the deficiency of constitutional enforcement against the President is well documented,88 the deficiency in the police context cuts against assumptions underlying the distinction between war and crime. In the criminal context, it could be argued that even though the standard for criminal liability is similarly weak for the President and the police, the real threat of prosecution makes it more constraining for the police. But the reality of criminal prosecution of lethal force does not support such a claim. While there is insufficient data to clearly establish the deterrent effect of criminal law on the use of lethal force, there is reason to believe it is not strong. First, charges and convictions are infrequent. Of the thousands of fatal shootings by police from 2005 to 2016, only fifty-four officers were charged, and eleven convicted.89 The weak substantive standard accounts for some of this, but there are also institutional challenges in getting to a conviction such as the role of police unions in opposing convictions,90 the conflict of interest presented by the relationship between prosecutors and police departments,91 and the power of grand juries to not indict, even where probable cause is found.92 Finally, even if the threat of a successful prosecution were serious, the deterrent value is weakened in the context of lethal force, because in the instances where officers do perceive a threat to their lives (reasonably or not), it is unlikely that the threat of prosecution will alter their decisionmaking. When deciding between loss of life and litigation, the preference is unsurprising. To be sure, criminal liability is more possible for police than for the President, but it would be a mistake to think that the mere possibility is meaningfully constraining. Constitutional litigation and criminal prosecution are the visible centerpieces of police lethal force regulation. Citizens expect them to work and protest for their efficacy.93 Scholars invest time and research to improve the constitutional standard and enable effective criminal prosecutions. Practitioners litigate toward these ends. Investment in these vehicles of regulation, however, is not clearly efficient. This is evident from how little regulation the standards provide. Even while constitutional and criminal law are useful regulatory tools in other areas, advocates might do well to view them with skepticism for the regulation of lethal force.

III. Overperforming War Paradigm Regulatory Tools

In regulating the President’s actions abroad, great hope has never rested on finding criminal or constitutional liability. Instead, an alternative infrastructure of regulation has emerged to constrain the President. This regulatory infrastructure rests primarily on three tools: international law, authorizing statutes, and internal policies. These tools also regulate police lethal force. The difference is that for the Executive, they play a central role and impose meaningful restrictions; for the police, they are background players that add little to the weak constitutional standard. While these tools are frequently understood as second best alternatives to the preferred tools of constitutional and criminal law, here in the context of lethal force they prove to be more effective. The following sections analyze these regulatory tools in turn, showing how they constrain the President more than the police.

A. International Law

International law exerts influence on lethal force decisions made by the President, but not on those made by the police. This is unfortunate because international law offers a functional approach much needed for all lethal force regulation. International law does not rely exclusively on legal sanction for compliance, but also leverages the force of moral and normative sanction.94 The modes of sanction in international law are diverse because there is recognition that judicial sanction cannot bear the compliance burden alone. International humanitarian law (IHL), for example, acknowledged this limitation in its design. Because it could not be presumed that courts would enforce IHL, its development “proceeded on the basis of moral sociology, discerning the possibility of a viable norm of restraint.”95 The product of such an effort, at least with respect to lethal force, is a set of analytically useful limitations for the President that unfortunately do not apply to the police.

1. The Police

International law plays functionally no role in the regulation of police lethal force within the United States. United Nations Basic Principles do establish limits on the use of force by law enforcement officials. For example, Principle 9 dictates, inter alia, that lethal force can only be used when “less extreme means are insufficient.”96 Basic principles, however, are non-binding, and perhaps more importantly from a practical lens, have little norm-forming power to regulate domestic law enforcement policy.

2. The President

International law, by contrast, does play a role in the regulation of executive lethal force. While it cannot be said that the President expects enforcement of international law against her decisions to use lethal force, it is evident that the norm-forming component of international law has influenced executive branch decisions to use lethal force. The OLC memo analyzing the Awlaki strike dedicated significant attention to international law (much more than to the Fourth Amendment, for example).97 It first established that the killing of Awlaki sat within a non-international armed conflict, which made international humanitarian law the relevant legal framework.98 It then listed IHL’s “four fundamental principles that are inherent to all targeting decisions”: military necessity, humanity, proportionality, and distinction.99

Some of these principles supplement the Fourth Amendment requirements in meaningful ways. Proportionality, for example, requires a comparison between the “concrete and direct military advantage” and the expected “incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof.”100 Notably, this is a balancing framework. But because it requires articulation of the costs and benefits, it is stronger than Fourth Amendment “reasonableness” balancing.101 There is some evidence that this constraint played a functional role in decisions leading up to Awlaki’s death—the strike was delayed for around a month while Awlaki was in a village, so that he could be killed on a road far from populated areas.102

Humanity is another principle that requires more than the Fourth Amendment. Then-Attorney General Eric Holder described the principle of humanity as requiring that the use of force “will not inflict unnecessary suffering.”103 This adds a layer of analysis that is doctrinally murky in the Fourth Amendment. While the type of force used (car, bullet, chokehold) is likely considered in Fourth Amendment reasonableness analysis, there is no doctrinal requirement to consider whether less pain-inducing means are available. Proportionality and humanity, then, provide an analytical strength to executive lethal force analysis that is absent from the police context.

Other IHL principles, however, do not add much to the Fourth Amendment standard, or otherwise are not relevant to the police context. The principle of necessity requires that the target, “have definite military value.”104 The equivalent in the police context is the requirement (per Garner) or factor (per Graham) that the suspect poses a danger. Separately, the principle of distinction requires “reasonable certainty” that the person the missile is trained on is actually the person on the kill list.105 While this is of some significance in the drone context, it adds little when placed in the policing context. It is an implicit requirement of the Fourth Amendment that the force is used against the proper person.

While it would go too far to say that international law creates binding rules on the Executive to which the police are not subject, IHL does provide analytic and norm-forming tools for lethal force analysis that are non-existent in the police context. International law is characterized as a regulatory “middle road”106 because even while it allows “[b]elligerent [p]arties much leeway,” it also circumscribes their “freedom of action.”107 This functionalist approach of international law might be useful to domestic police use of force.108

B. Authorizing Statutes

For both the President and the police, authorizing statutes could meaningfully limit the circumstances under which force is permitted. Because authorizing statutes involve the creation of limits through the ex ante legislative process, rather than the ex post judicial process, they are better suited to formulate specific and informed limits on lethal force.109 While authorizing statutes are largely impotent in the police context, they do provide some limits on executive lethal force.

1. The Police

Police authorizing statutes are often vague and permissive.110 New York City’s is typical, authorizing the police to, inter alia, “preserve the public peace,” “prevent crime,” and “detect and arrest offenders.”111 The breadth diminishes capacity for constraint. While some jurisdictions have passed legislation that specifically limits certain areas of police practice, such as the push to statutorily require the use of body cameras, such efforts are largely absent in the context of lethal force.112 Authorizing statutes effectively forfeit the lethal force regulatory game.

2. The President

The statutory authorization of executive lethal force is more complicated and more limiting. Executive authority to use lethal force is analyzed under the Youngstown framework, such that authority is conditioned by the relationship between statutory permission and constitutional power.113 This framework provides more than one route to find authorization for lethal force outside of active hostilities. Both President Bush and President Obama, for example, have claimed inherent Article II powers to use force even without congressional authorization.114 Whether that inherent power would extend to the killing of Awlaki is unsettled, but also presently moot, because OLC has found that such a strike is authorized by statute. The executive branch exercises lethal force either through the Department of Defense (DOD) or the Central Intelligence Agency (CIA), each of which has distinct statutory authority. DOD operations constitute “military activities” which are governed by statutes in Title 10 of the U.S. Code. CIA operations, by contrast, constitute “intelligence activities” which are governed primarily by the National Security Act of 1947 (NSA).115 Prior to the killing of Awlaki, it was undecided which agency would take the action, so OLC analyzed the authorizing statutes of each, and determined that both granted sufficient authority.116 The relevant question here, however, is how and whether the statutes constrain the use of force. As for DOD authority, the Authorization for the Use of Military Force (AUMF) authorizes the President,

to use all necessary and appropriate force against those nations, organizations,    or persons he determines planned, authorized, committed, or aided the terrorist attacks that occurred on September 11, 2001, or harbored such organizations or persons, in order to prevent any future acts of international terrorism against the United States by such nations, organizations or persons.117

The key limitation here, absent in the police context, is the anchoring to 9/11.118 While an equivalent in the police context is not obvious, we might imagine an authorizing statute that only permitted lethal force in response to certain serious crimes.119 Such a component of an authorizing statute would limit constitutional doctrine, which now permits lethal force in high-speed car chases that escalate from mere traffic violations. To be sure, the AUMF does not live up to the potential regulatory control of an authorizing statute, but it does provide more limits than does the average police authorizing statute.

CIA authority comes from a different statute—the NSA.120 The first four categories of authority granted to the CIA by the NSA enable it to centralize and streamline traditional intelligence activity. The “fifth function,” though, is a catchall clause, which now reads, “[t]he Director of the Central Intelligence Agency shall . . . perform such other functions and duties related to intelligence affecting the national security as the President or the Director of National Intelligence may direct.”121 It is this phrase that empowers the CIA to take covert action internationally.122 The assassination attempt of Fidel Castro, for example, was a product of the fifth function.123 While OLC redacted its analysis of how this authorized the strike against Awlaki, it nonetheless concluded that it did.124

One limitation likely included in OLC’s fifth function analysis is that most individual covert actions require approval through an executive finding, whereby the executive branch determines whether the action “is necessary to support identifiable foreign policy objectives” and “is important to the national security of the United States.”125 Frankly this is not a difficult hurdle, but it is one step of review and reflection not required of the police. What this might look like in the police context is an authorizing statute that required police departments to produce a lethal force policy—something that some departments still lack.126

In sum, while the authorizing statutes of executive lethal force are not models of restriction, they do provide small mechanisms of constraint not included in police authorizing statutes. One plausible explanation for this divergence is that members of Congress understand that if they fail to limit the Executive through authorizing statutes, little regulation will appear elsewhere. By contrast, such a burden is not felt in the police context where the regulatory work is expected to be done by the courts.127

C. Internal Policy

Self-regulation by the police and the President provides certain comparative advantages. The police and the President have the most intimate knowledge of the quality of the threat to which they respond. As such, to the extent they are willing, they can impose specific, clear, and functional guidelines on how to avoid situations that tend to result in the use of lethal force. Moreover, making these rules public would improve accountability.128 This mode of regulation is utilized more by the President than by the police. While White House lethal force policy takes the regulatory system of constitutional, statutory, and international law as a floor and builds up from there, police departments on balance do not add to the underlying criminal and constitutional requirements.

1. The President

At least since 9/11, internal policy on the use of lethal force abroad was developing within the White House. It became public in a piecemeal fashion, the most comprehensive version of it being released in a 2013 document called “U.S. Policy Standards and Procedures for the Use of Force in Counterterrorism Operations Outside the United States and Areas of Active Hostilities.”129 The document claims that it includes “certain key elements” of the standard and also that “[administration officials] are continually working to refine, clarify, and strengthen [the] standards and processes.”130 The document contains a number of policy restrictions on the use of force that go beyond the requirements of law. First, the policy states that one requirement for the use of force is an assessment that “no other reasonable alternatives exist to effectively address the threat to U.S. persons.”131 This incorporates two limits not required by law. First, it invokes a principle of force minimization requiring that non-lethal or less-lethal force be used wherever possible. As noted earlier, constitutional and statutory law do not require that less-lethal force be used wherever possible. Second, the language in the policy also implicitly requires that non-violent solutions, such as de-escalation, be considered before the use of force. Constitutional and statutory law do not require this. In Scott, for example, the Court conceded that the threat could have been addressed if the cops simply stopped chasing the fleeing car. Nonetheless, the Court found that the police decision to ram the car was reasonable under the Fourth Amendment.132 The White House policy is also clear about the imminence requirement in a way that the constitutional standard is not.133 The policy states that force may only be used against a target that poses a “continued and imminent” threat.134 Compare this to Graham, where the imminence and timing of the threat is relevant but not dispositive.135 Even the Garner requirement, slightly stronger than Graham’s, leaves open the possibility that the severity of the crime committed may justify the use of lethal force even when the imminence of a future threat is questionable.136 The executive policy explicitly prevents such an ambiguity by stating that “[i]t is simply not the case that all terrorists pose a continuing, imminent threat to U.S. persons; if a terrorist does not pose such a threat, the United States will not use lethal force.”137 Another requirement added by the policy is near certainty that non-combatants will not be injured or killed.138 This is absent from statutory and constitutional law. Consider Plumhoff v. Rickard, where the Court found that the presence of a non-active passenger—who was killed by the decision to ram the driver’s vehicle—was irrelevant to the reasonableness of the use of force.139 Finally, the policy requires that the decision to use lethal force against a citizen must be made at the “most senior levels” of U.S. government.140 As noted above, the OLC memo found that this review was sufficient for the Fifth Amendment, but did not state whether it was necessary.141 The import of the policy, then, is to make it necessary. While one might doubt the feasibility of having every instance of police lethal force approved by senior police officials, it is sometimes possible. Consider for example the killing of Micah Johnson, where the decision to detonate a robotic bomb was not a quick one.142 A reasonable rule might require that where time permits, officers must get approval. In sum, even though not demanded by statute or the constitution, the executive policy requires that force only be used as a last resort, that only the minimum required force is used, and that the consequence of lethal force to bystanders is considered.143 It also creates a strict imminence requirement.144 In this way, executive branch policy supplements existing legal requirements.

2. The Police

Police policy does not. Note first that while there is one White House, there are close to 18,000 policing agencies in the U.S.145 The task then cannot be to compare each agency policy to the White House’s lethal force policy. Instead, the comparative question is necessarily a rough one, asking whether the average police department policy imposes limits on lethal force not already required by law. Even this, however, is difficult to establish. A recent study surveyed police departments for more detailed information about their lethal force policies.146 While this survey does not use random sampling or get close to catching most of the nearly 18,000 agencies, it focuses on the fifty largest police departments as a proxy. The authors of the survey intentionally chose the largest departments because they expected them to have the most developed, researched, and detailed policies on lethal force. In light of this expectation, the results of the survey are troubling. First, there is a problem with transparency and accountability. Unlike the public nature of the White House policy, the study found that only seventeen of the fifty largest agencies make their policies available publicly.147 More substantively, the survey found that most of the departments surveyed did not discuss de-escalation tactics in their policies.148 Neither did most policies have requirements to minimize the degree of force used.149 Compare this to the executive requirement that no alternative exists to the use of force.150 Additionally, only a slim majority of the policies encouraged or required verbal warnings before the use of force.151 Therefore, even among the policies expected to be most thorough, only some added to the constitutional standard.152 Internal policy has a meaningful role to play in lethal force regulation. It allows for the actors with the most intimate knowledge of the dynamics of lethal force to self-impose specific rules and make them public for the sake of accountability. Currently, the executive branch makes use of this tool but police departments do not. Why exactly this divergence exists is debatable, but one reasonable explanation is the fact that constitutional litigation is available against the police but not the President. The presence of constitutional litigation both symbolically deflects responsibility from police departments to create strict rules of their own and rationally disincentivizes such policies as those policies can then be used to hold police departments liable. In this way, not only does constitutional litigation fail in its own right, it also might crowd out the creation of alternative modes of regulation, such as internal policy.

IV. Lessons for Police Constraint

To the extent that war paradigm regulation—i.e. international law, authorizing statutes, and internal policy—has produced functional constraints on the President, advocates of police reform may find it productive to use similar types of regulation to guide the police. This section offers suggestions on how this might be done.

A. Demoting Ex Post Regulation

First, the relative constraint of war paradigm regulation of lethal force calls for humility about the regulatory capacity of criminal and constitutional law. Not only is it true that criminal and constitutional law are structurally suboptimal as regulators of lethal force, it is also possible that by failing to recognize this fact, advocates for police reform are channeling energy into the wrong places. We should recognize the possibility that investment in constitutional and criminal regulation might crowd out investment in alternative infrastructures of regulation.

To reflect this recognition, discourse in the aftermath of tragic incidents of lethal force could assert clearly both the value and the shortcomings of constitutional and criminal law as a response. The value of criminal and constitutional law after incidents of lethal force is largely about accountability. As a reflection of a moral imperative against unjustified police violence, accountability is important and individuals and communities are not misguided in pursuing it. That said, accountability should not be confused with functional regulation. As this Article has demonstrated, criminal and constitutional law do not provide sufficient constraint on police use of force, and that is not likely to change—the difficulty of second guessing officers who assert a fear for their lives is inherent to ex post enforcement. Therefore, demands that are made by advocates could include not only legal accountability, but also steps to enact ex ante regulation.

Hopefully, a public recognition that ex post criminal and constitutional enforcement serves accountability goals but not regulatory goals will generate an investment in alternative regulatory tools. Once free from the illusion that criminal and constitutional law suffice to constrain the police, finding other tools of constraint will become a more urgent project.

B. Promoting Ex Ante Regulation

One place to look is the regulation of executive lethal force. There, the regulatory gap left by criminal and constitutional law is recognized, and international law, authorizing statutes, and internal policy are used to fill it. Numerous tools of constraint used in the executive context may be drawn upon by reform advocates for application to the police. First, the theoretical underpinnings of regulation in the international context may be helpful for discussions about developing functional police regulation. In particular, international humanitarian law, in recognition that it may not be given teeth by the courts, was designed in consideration of workable constraints that could be enacted through norm formation.153 A similar process—discerning viable norms of restraint—is a useful starting point for all discussions of police policy.

Moving from there, a number of specific rules imposed on the Executive may be productively applied to the police through authorizing statutes and police policies. One such rule is that force must only be used as a last resort.154 If incorporated into a police policy, this would instruct officers to use other methods to mitigate the threat where possible. Such a policy would benefit from accompanying training on de-escalation, as those tactics would very often be the reasonable alternatives to lethal force. The case of Scott illuminates the change that this rule would make.155 There, an officer chose to ram the back of a fleeing car instead of letting it go—two methods to end the present risk of a high-speed chase. The last resort rule would require that the officer let the car go instead of ramming it. To be sure, some might object to the incentive structure this sets up, but it is no different than the approach the Court took in Garner, instructing officers not to shoot at non-dangerous fleeing felons.

Another candidate for translation is the force-minimization principle. Such a principle is reflected in international humanitarian law, requiring that the use of weapons does not “inflict unnecessary suffering.”156 Applied to the police context, this rule would control the type of force used and demand that officers use the minimum force required so as not to inflict unnecessary suffering. While consideration of these factors is generally baked into current Fourth Amendment reasonableness analysis, there is no rule requiring minimization.

Police policies should also establish clarity on the rules surrounding the presence of non-active parties or bystanders. In the executive context, it is required that the consequence of lethal force to bystanders is considered.157 If translated to a police policy, this rule would very likely save lives. In the situation that led to Plumhoff v. Rickard, for example, where an officer rammed the driver’s vehicle and in doing so killed the passenger,158 a police policy consistent with the executive policy would have changed the outcome. Had the officer been aware of internal policy requiring near certainty that non-active people would not be injured or killed, perhaps the officer would not have rammed the vehicle.

Police policies might also be aided by a clear imminence requirement. While the imminence requirement for police was left unclear by Garner,159 the executive policy states that force may only be used against a “continuing and imminent” threat and that “[i]t is simply not the case that all terrorists pose a continuing, imminent threat to U.S. persons.”160 Translated to the police context, this language might express that an “imminent threat cannot be inferred solely by the fact that somebody previously committed a crime.” Such language would clarify the ambiguity left open by Garner.

In sum, advocates for police reform can gain much by looking to regulation of the President’s use of lethal force. First is a paradigm-shift in how we think about the regulation—moving from a reliance on the ex post enforcement of criminal and constitutional law to a call for ex ante regulations that instruct officers before violence occurs. Second is the adoption of specific constraints that have been leveraged in the executive context, but not in the police context, including the last resort rule, the principle of force minimization, the bystander rule, and the imminence rule.

V. Conclusion

The crime/war distinction forecasts that police lethal force will be more constrained than executive lethal force. But the reverse is true. The regulation of lethal force by the police and the President is similar in the way that criminal law and constitutional law fail to provide meaningful constraint. The difference is that while the President is also constrained by alternative forms of regulation, the police are not. Indeed, while both are largely free from constraint by the traditionally “first best” tools of regulation, only the President is constrained by the “second best” tools of international law, authorizing statutes, and internal policy. Recognizing this relationship generates two further claims.

First, it complicates the assumption of the crime/war distinction that rights are stronger and that the government is more limited at home. The presence of criminal and constitutional judicial proceedings in the police context (as compared to the absence in the executive context) triggers an expectation of rights protection. This Article determines that this expectation is unwarranted. And while this Article presents the law of lethal force as merely one exception to the utility of the crime/war distinction, it is worth asking if there are others. One possibility is government surveillance. The crime/war distinction was central to the purpose of FISA, which sought to selectively relax restrictions on government surveillance.161 Approval for warrantless surveillance under FISA required, inter alia, a showing that the primary purpose of the surveillance was to gather foreign intelligence information.162 Surveillance would not be approved by a FISA court if the primary purpose was criminal law enforcement.163 That distinction no longer exists. Capping many years of tension over the distinction, including failed attempts by Congress to diminish it with the PATRIOT Act, the Foreign Intelligence Surveillance Court of Review eradicated it in 2002.164 The court held that gathering foreign intelligence information need only be a significant purpose of the surveillance,165 in part because the crime/war distinction was “inherently unstable, unrealistic, and confusing.”166 The result is that where there once was a consequential distinction between rights against surveillance in the criminal context and the war context, there no longer is.167 Perhaps, then, this is another example where the crime/war distinction is unhelpful, putting more pressure on its utility.

Second, the descriptive claim of this Article also lends support to growing calls for ex ante regulation of the police. By noting that regulation of executive lethal force is aided by ex ante tools, this Article suggests that regulation of police lethal force could be too. But more than that, it advances more narrow claims in two directions. The first is that lethal force, as compared to policing in general, is in especially urgent need of ex ante regulation. This is because the problems of ex post regulation are particularly acute when applied to the qualities of lethal force. Consider, for example, how deterrence, judicial decision-making, and remedies operate in the context of lethal force. The regulatory worth of ex post regulation rests on deterrence, yet the value of deterrence is weakened in the context of lethal force because a government actor is operating in response to what she perceives to be life-threatening. This same quality—that government actors are responding to perceived threats to life—also complicates judicial decision-making. If judges are hesitant to exclude evidence in Fourth Amendment search jurisprudence, consider the extent to which they would be hesitant to impose sanctions on someone claiming to have chosen between individual or collective death and the use of lethal force. Add to that the factual and evidentiary complexity of lethal force incidents, which makes it difficult for judges to second guess claims of existential threat. This is not to excuse or condemn judges, but simply to point to the structural forces that push them away from finding liability. Finally, lethal force is lethal and no remedy can resuscitate. Whereas monetary, injunctive, or evidentiary relief may be meaningful in certain situations, the finality of lethal force renders them symbolic at best. These qualities of lethal force make it impervious to ex post criminal and constitutional regulation. Accordingly, ex ante guidance for lethal force is especially warranted.

The second is that a misguided reliance on criminal and constitutional law is crowding out alternative infrastructures of regulation in the police context.168 This dynamic is illustrated by the public response to lethal force in each context. For example, when reports circulated that President Trump intended to relax internal constraints on lethal force,169 critics responded with calls to constrain him with Congressional action170 and international norms.171 Contrast this response to that typically following incidents of lethal force by the police: a call to prosecute and sue using criminal and constitutional law.172 While crowding out is difficult to prove as a matter of fact, this comparison indicates that a belief in the efficacy of criminal and constitutional law for the police might be hindering the development of other regulation. This raises the question, what would we demand of police departments if we did not believe that criminal law and constitutional law already put demands on them? One answer might be in the executive context, where a sober recognition of the weaknesses of constitutional and criminal law results in an investment in other tools, such as international law, authorizing statutes and internal policy. Rather than doubling down on criminal and constitutional law, regulation for police lethal should follow suit.

A first step to making this happen is extricating police lethal force from the assumptions of regulation in the crime paradigm. This Article intends to do that. Despite the forecasting of the crime/war distinction, our rights are no stronger against the police’s Sig Sauer in Missouri than against the President’s Hellfire missile in Yemen. Discourse, scholarship, and the law would do well to acknowledge this fact.

Suggested Reading