CHAPTER 6. The Fight Against Terrorism
(From: LAW, ECONOMICS, AND MORALITY, by Eyal Zamir and Barak Medina,
forthcoming Oxford University Press)
This is a chapter of our forthcoming book, titled Law, Economics, and Morality, in which we propose to integrate threshold deontological constraints (and options) with cost-benefit analysis (CBA), thus combining economic methodology with deontological morality.
The chapter presents a constrained CBA of measures taken in the fight against terrorism. It begins by characterizing and criticizing existing normative economic analysis of the fight against terrorism as reflecting a simplified ad-hoc balancing. It then presents the central deontological constraints pertaining to the fight on terror. The bulk of the chapter discusses threshold functions that should be employed to determine the permissibility of such measures as targeted killings and torture. We discuss the factors affecting the evaluation of the act’s relevant net benefit, and those determining the amount of net benefit required to justify an infringement.
We argue that standard economic analysis fails to take into account critical distinctions. These include the distinction between different goals of antiterrorist measures (including retribution, deterrence, preemption and pressure); the difference between harms the state inflicts through antiterrorist measures, and those resulting from unthwarted terrorist attacks; and the distinction between intended and unintended harm. Deontologically-constrained CBA, which incorporates all of these distinctions, is shown to be methodically workable and normatively superior.
Chapter 6. The Fight Against Terrorism
B. Economic Analysis of the Fight Against Terrorism
C. The Constraint Against Harming Persons and the Fight Against Terrorism
1. General Considerations
2. Harming Aggressors as a Constraint Infringement
D. Constrained Economic Analysis of Intended Harm
2. Goals of Anti-Terrorist Measures: Preemption, Retribution, Deterrence, and Pressure
3. Basic Elements of the Threshold Function
4. The Net Benefit
(a) The Relevant Variables
(b) Marginal Net Benefit and Alternative Courses of Action
5. The Threshold
(b) Probability of the Terrorist Attack
(c) The Aggressor’s Culpability
E. Constrained Economic Analysis of Unintended Harm
2. Constructing the Threshold Function
3. Killing Persons Who Are Doomed
4. Victims’ Moral Responsibility and Nationality
F. Measures Involving Both Intended and Unintended Harm
CHAPTER 6. The Fight Against Terrorism
The fight against terrorism poses significant moral and legal challenges to liberal democracies. The attempts of potential terrorists to disguise themselves among innocent, civilian population, as well as the substantial damage created by acts of terror, present policy-makers with tragic choices. For instance: What type of preemptive measures can a state legitimately take against persons before proving their culpability—or even their evil intentions—in a court of law? Is it legitimate to save lives by harming dangerous individuals through extra-judicial methods, such as targeted killings, administrative detentions, or aggressive interrogation techniques? Is it justified to inflict harm on innocent persons in an attempt to save others?
The extensive literature evaluating the reaction of the democratic world to the terrible events of September 11, 2001 largely reflects a division between two schools of thoughts. One camp, which mostly criticizes the United States‘ reaction, rests its arguments on ‘principles.’ According to this view, the Constitution “represents a collective commitment to principles ... [and] the recognition that ‘pragmatic’ cost-benefit decisions will often appear in the short term to favor actions that may turn out in the long term to be contrary to our own best principles.”1 The other camp, to which legal economists usually belong, rejects precisely this commitment to principles. It claims that there is no consensus on the content of these principles, as “text, tradition, precedent, and reason so often tug in different directions.”2 Consequently, “[r]ealism requires recognition that constitutional decision-making…is driven in the main by policy judgments,”3 which are based exclusively on estimating the costs and benefits associated with each government practice. This chapter questions the alleged dichotomy between principles on the one hand, and cost-benefit analysis on the other, by demonstrating the plausibility and desirability of incorporating deontological constraints (‘principles’) with economic analysis (’balancing’).4
We begin, in part B, by briefly characterizing and criticizing existing normative economic analyses of the fight against terrorism as reflecting a simplified ad-hoc balance of interests (or act-consequentialist) approach. Part C presents the central deontological constraints pertinent to the fight against terrorism. We focus on the constraint against actively/intentionally inflicting harm. Additionally, we address the debate as to whether an aggressor retains his right to life regardless of his actions or evil intentions, and suggest that preempting an aggressor should also be considered as a constraint violation, at least in cases where the guilt of the harmed person has not been established by fair judicial proceedings. Parts D and E discuss the threshold functions that should apply when determining the permissibility of inflicting intended and unintended harm, respectively. We discuss the factors that should be considered in evaluating the act’s relevant net benefit, and those determining the size of net benefit required to justify an infringement. Part F analyzes anti-terrorist measures that involve inflicting both intended and unintended harm, demonstrating an integration of the two types of threshold functions.
B. Economic Analysis of the Fight Against Terrorism
Terrorism is conventionally defined as the use of violence causing death, great bodily harm, or serious psychological damage to innocent individuals with the intent to cause those harms or with wanton disregard for them, for the purpose of coercing or intimidating a specific group or government or otherwise gaining some perceived political or military benefit.5 Compared to other criminals, terrorists are often willing to sacrifice their lives for their cause, which significantly diminishes the deterrence effect of legal sanctions on terrorism. Additional distinctive features of terrorism are its highly organized form and the extensive indirect adverse effects of the threat of terror. A substantial portion of the economic analysis of terror is devoted to a positive, rather than normative, inquiry of these distinctive characteristics of terrorism.6 Such positive analyses include theoretical and econometric evaluations of the efficacy of the use of various anti-terror policies;7 political-economy assessments of the characteristics and sources of terrorism;8 and empirical studies of the adverse economic effects of terror.9
These aspects of the economic analysis of terror and the fight against it are important in evaluating existing security policies and in designing new ones. Our focus is, however, on the branch of economic analysis that aims to provide normative evaluation of such policies. Two notable contributions to this branch of inquiry are Richard Posner’s Not a Suicide Pact: The Constitution in a Time of National Emergency, and Eric Posner and Adrian Vermeule’s Terror in the Balance: Security, Liberty, and the Courts.10
The normative economic analysis of the fight against terrorism advocates an ad-hoc balance of interests approach. It opposes the counterintuitive claims of absolutist deontology,11 and holds that the use of a security measure is justified whenever its expected social benefits exceed its costs. As succinctly stated by Richard Posner, “one is not to ask whether liberty is more or less important than safety. One is to ask whether a particular security measure harms liberty more or less than it promotes safety.”12 This approach rejects the notion of civil liberties as constraints on consequentialist balancing. It claims that “security and liberty trade off against one another,”13 and thus “governments should, and do, balance civil liberties and security at all times.”14
The concept of ‘balancing’ is, of course, not unique to the economic approach. Courts frequently employ balancing tests when there are two or more competing interests, construed, for instance, as a conflict between an individual right and state need.15 However, the process of balancing advocated by the economic approach is distinctive in several respects. For one thing, CBA weighs competing interests in monetary terms, reflecting people’s preferences.16 Another singular aspect is the use of ad-hoc balancing, rather than a so-called ‘definitional’ one. The balancing endorsed by standard economic analysis is ad-hoc not only in the sense that it prefers a case-by-case inquiry to a global one,17 but also in that it balances the relevant interests, rather than competing constitutional principles or values. For instance, when the United States Supreme Court used the terminology of balancing in analyzing the legitimacy of detaining suspected terrorists in the case of Hamdi v. Rumsfeld, it referred to “[s]triking the proper constitutional balance” between the security needs and “the values that this country holds dear…[such as the] commitment to due process.”18 In contrast, notwithstanding the rhetoric of comparing liberty and security, economic analysts take into account only the concrete individual interests that are at stake, rather than abstract values and principles. In this sense, they are act-consequentialists.
Posner and Vermeule concede that in principle, one may apply what they term a “non-welfarist” balancing, in which “the effects of actions on rights are themselves among the consequences to be evaluated.”19 Treating right infringements as a cost does not transcend consequentialism, yet the adoption of such a more sophisticated theory of the good may somewhat narrow the gap between consequentialism and deontology.20 At any rate, they do not adopt such an approach. The economic analysis of the fight against terrorism thus does not consider the mere infringement of a basic liberty or some other fundamental principle as a social cost in the constitutional calculus. What counts is only the policy’s adverse effect on the welfare of individuals. Under this view, the content of a concept such as liberty is nothing but the outcome of the balancing process, as it is determined exclusively by the scope of society’s existing needs: “[it] is the point of balance [that] determines the optimal scope of the right.”21 Liberty is not perceived as a normative concept, but rather as a descriptive one.
Posner and Vermeule, as well as Richard Posner, oppose setting any deontological constraint that would limit states in taking security measures in the fight against terrorism, for two reasons. First, as far as the constraints are based on second-order institutional and empirical concerns (rule-consequentialism), they argue that these concerns are highly exaggerated and unsubstantiated.22 In fact, they suggest that the main concern is that deontological constraints might “block government’s attempts to adjust the balance as threats wax and wane.”23 The real danger, so they argue, is that “civil libertarian panic about the specter of authoritarianism” will hinder cost-justified security measures.24 Thus, they not only reject the view of rights as trumps but also reject the position that one should “place a thumb on the scale” to assure that the balance struck in any particular situation properly reflects the central position of the relevant liberty in the constitutional scheme. If anything, they place the thumb on the side of security.25 Second, Posner argues that deontological constraints, including those recognized as constitutional principles, are indeterminate and subjective, as they reflect each decision-maker’s moral and religious values and personal life experience.26 Posner and Vermeule suggest that even though incorporating deontological constraints and setting thresholds can in principle be justified on the basis of first-order moral considerations, this approach is inapplicable, since setting the threshold comes at the price of “reduced theoretical coherence and an arbitrary flavor.”27 Consequently, according to standard economic analysis, to justify a security measure, it is not necessary to show that it would generate sufficiently high social benefits to override a deontological constraint. It suffices that the benefit exceeds the cost of the policy’s direct adverse effects.
In accordance with standard act-consequentialism, the economic analysis disregards the doing/allowing distinction. Hence, it equates the active infliction of harm by security measures with not preventing a comparable harm to the potential victims of terrorist activities.28 Standard CBA also rejects the distinction between saving or harming unidentified (‘statistical’) people and saving or harming identified ones.29 As a result, human rights and liberty are placed on both sides of the scale, thereby offsetting the status of basic liberties in the process of balancing.
Legal economists further propose to replace an actual, detailed balance of interests with a hypothetical one, referring to some baseline of normalcy. According to this notion, the state always operates at a ‘Pareto frontier’ with regard to the balance of security and liberty; and hence there is a perfect trade-off between ‘liberty’ and ‘security.’ Any increase in the demand for security, resulting from the threat of terror, entails a decrease in liberty.30 It follows—without even having to conduct a detailed CBA—that infringements of deontological constraints are a necessary and justified corollary of the fight against terrorism.
CBA allows legal economists to argue that judges should defer to executive decisions, and that judicial review of governmental action should be relaxed or even suspended in times of emergency.31 This position rests on the concern that judges have limited institutional capacity and knowledge needed to evaluate security policies.32 Once one assumes that designing anti-terrorist policies entails no principled normative deliberation but merely an aggregation of costs and benefits, governmental agencies seem better qualified than courts to perform the task of evaluating security policies.
Not surprisingly, standard normative economic analysis of the fight against terrorism legitimizes a rather broad range of measures. It endorses policies such as indefinite detentions, warrantless domestic wiretapping, and coercive interrogations.33 It justifies restrictive measures against persons whenever the expected benefit of those measures exceeds their costs, even if the probability that these persons would have been involved in terror is marginal. For example, Richard Posner concedes that even in radical Islamist communities, only a small percentage of people are willing to commit, or even abet, terrorist acts. He nevertheless argues that since there are millions of Muslims in the United States, and since a few terrorists may cause catastrophic harms, the government may take measures to curtail the freedom of speech of radical imams.34 Posner and Vermeule even question the desirability of the international laws of war. They contend that “if the laws of war are intrinsically bad, the United States should not only violate those that interfere with the war on terrorism; it should advocate the abolition of all laws of war.”35
For the reasons elaborated in the first part of this book, we find both the consequentialist underpinnings of standard normative economic analysis of the fight against terrorism and much of its implications unacceptable. The claim that a state’s failure to effectively eliminate threats of terrorism is morally equivalent to actively/intentionally harming people to prevent such threats disregards the doing/allowing and intending/foreseeing distinctions, and fails to distinguish between deontological options and deontological constraints.36
Furthermore, the claim that the risk of terror necessarily justifies losses in liberty is unfounded. The alleged tradeoff rests on questionable empirical assumptions. It assumes not only that at the baseline the state operates at a Pareto frontier with regard to the balance of security and liberty, but also that all background factors, such as the resources allocated to tackle the security threat, do not change in the move from normalcy to emergency. These assumptions are implausible.37 More fundamentally, the claim that there is a perfect trade-off between ‘liberty’ and ‘security’ treats the curtailment of liberty as a mere shorthand for reduction in people’s welfare. Once liberty is conceived as entailing a deontological constraint on actively/intentionally taking peoples’ lives or harming their bodily integrity and dignity, then two fundamental modifications of the balancing process are needed. First, to justify the infringement of constraints, the expected good outcomes (or the avoided bad ones) must meet a relatively high threshold. Second, only certain types of benefits and costs may legitimately be taken into account in this context.38 The following discussion describes the central characteristics of such a modified balance between liberty and security.
C. the Constraint Against Harming Persons and the Fight Against Terrorism
1. General Considerations
Anti-terrorist activities may entail infringements of any conceivable deontological constraint, including the constraints against deception and breaking promises. However, such activities primarily infringe the most basic deontological constraints; those against actively/intentionally killing people, injuring them, curtailing their freedom of movement and speech, and harming their human dignity.
Identifying the pertinent constraints and setting thresholds for the permissibility of their infringement rest on the ordinary deontological distinctions. These include the distinction between doing harm and merely allowing it. They also include the distinction between intending harm (which may be immoral even if the harm is merely allowed), and foreseeing harm (which is not necessarily immoral), and the related distinction between harming a person as a side-effect of attaining a legitimate result, and using a person as a means to aiding or saving others. A harm is intended even if the actor has no interest in imposing the harm except as a means to achieve some end. For instance, if the government puts pressure on a terrorist by harming innocent persons who are dear to him, the harm to these persons is intended, irrespective of the government’s lack of a direct interest in harming those persons. Harm is unintended only when it is a mere side-effect, that is, when the adverse outcome does not provide the actor with a reason for his behavior or an explanation for it.39
To illustrate, consider the following scenario:40 A state bombs a terrorists’ munitions plant, and as a result, kills civilians living nearby. Is the killing of the civilians intended? To answer this question, one must find out the aim of dropping bombs on the plant. One possibility is that the bombing was aimed at destroying the plant itself, damaging the terrorists’ capacity to build their own weapons, in order to prevent future terrorist attacks; another possibility is that the aim was to terrorize the civilian population supporting the terrorists, in order to break their morale and force them to surrender. Under the first possibility, the killing of civilians is a mere side-effect. Under the second, the state is intentionally harming civilians as a means to its end. Importantly, from the perspective of moderate deontology, in both cases the state infringes a constraint, yet the threshold that must be met to render the infringement permissible is different. The threshold is much higher in the second case, as intended harm reflects disrespect to human dignity. Victims have an interest not only in their fate but also in their status, in what can permissibly be done to them.41
2. Harming Aggressors as a Constraint Infringement
According to most deontological approaches, in evaluating the legitimacy of inflicting harm on a person, it is essential to take into account the person’s relevant moral responsibility. While consequentialism considers one’s fault only instrumentally, as harming the blameworthy is likely to deter future undesirable conduct, deontology plausibly ascribes intrinsic value to one’s fault. Imposing harm on a blameworthy person, at least in a context closely related to his blame, is less morally objectionable than imposing the same harm on an innocent person.
The German Constitutional Court referred to this issue in passim, when scrutinizing the constitutionality of a statute authorizing the shooting down of an aircraft intended to be wielded as a lethal weapon.42 The court decided that authorizing the military to shoot down a plane with innocent people on board is unconstitutional, as it infringes their right to dignity (even if these people are doomed to die anyway). However, the court held that if only terrorists are on board the aircraft, it is perfectly legitimate to shoot down the plane. The court reasoned that preempting the attackers does not infringe the moral constraint against killing people actively/intentionally because they forfeited their rights by their willful conduct.43
A moderate deontologist would agree that the blame of the individual harmed by the state is morally relevant in evaluating the act’s permissibility. However, contrary to the German court’s ruling, it is doubtful that inflicting harm on the guilty does not infringe any deontological constraint.
Consider, first, the imposition of criminal sanctions on a person convicted in criminal proceedings. Even if retribution obliges the state to impose the sanction,44 retribution can only justify some types of sanctions—those that are deemed “appropriate” to the severity of the person’s moral blame.45 A person’s guilt is not sufficient to establish the claim that sanctioning him does not infringe a moral constraint, since it is still required to show that the harm is justifiable in terms of the relevant theory of retribution. One’s wrongful actions thus do not negate all of one’s rights, and do not necessarily justify infringing the constraint against torturing or killing a convict.46 Moreover, retribution is not the only possible rationale for imposing criminal sanctions. Certainly, if one endorses other rationales for criminal sanctions, such as private and general deterrence, or incapacitation, the fact that imposing the sanctions is justified does not imply that it does not infringe the constraint against actively/intentionally inflicting harm.
The difficulties in characterizing the infliction of harm on the guilty as a non-infringement are significantly greater when—as is typically the case in the context of the fight against terrorism—no judicial determination of the person’s guilt precedes the harm. Even if retributive criminal sanctions do not infringe a constraint, it does not follow that the same is true of measures taken against suspected terrorists outside the criminal justice system. The requirement that retributive sanctions would follow a judicial finding of criminal liability reflects not only the concern of a wrong, and possibly biased, attribution of guilt in the absence of judicial proceedings, but also a normative judgment about fairness. A person can justifiably be subject to harm based on retribution only if his blame was determined in a way that meets the requirements of procedural justice.47
The more difficult case is that of inflicting harm with the aim of preempting an aggressor. John Locke famously argued that an aggressor forfeits his right to life.48 Following this view, several contemporary scholars endorse the so-called ‘forfeiture argument’: it is permissible to kill in self-defense, for a person possesses the right to life only so long as he does not pose an unjust, immediate threat to others.49 This position is subject to extensive and, in our view, convincing criticism.
The presumption that the aggressor actually chooses to forfeit his right to life is hard to sustain.50 Moreover, this presumption is incompatible with the notion that the right to life is inalienable.51 Thus, even if a would-be suicide-bomber intends to sacrifice his life in order to kill others, it does not follow that when determining what measures are legitimate to preempt the attack, one can assume that the would-be suicide-bomber does not possess a right to life. To the extent that the aggressor’s forfeiture of the right to life rests on his culpability or malice (and not merely on the threat he imposes), the argument fails to justify the right to self-defense against an innocent attacker. It is also subject to the difficulties discussed above, regarding the lack of fair judicial proceedings for establishing one’s guilt.
Most importantly for our purposes, the forfeiture argument does not quite explain the limitations imposed on the right to self-defense, such as the requirement that the preemption is necessary, and that the harm inflicted on the aggressor is proportionate to the harm that the aggressor aims to inflict. If the aggressor has no right to life, why impose these limitations?52 It is also unclear what types of constraints or rights the aggressor forfeits as a result of his aggression. Supporters of the forfeiture argument respond that the aggressor forfeits his right to life only at the point at which killing him is necessary and proportional.53 This response, however, is tantamount to arguing that a person has no right not to be subject to a justified infringement. It acknowledges that the aggression itself is insufficient to render the killing of the aggressor a non-infringement.
In sum, it seems that an aggressor retains his right to life regardless of his evil actions or intentions. The German court’s obiter dictum—that shooting down a plane when only the terrorists are on board would not infringe any moral constraint—is therefore unsupported. At least in cases where the guilt of the harmed person has not been established through fair judicial proceedings, any infliction of harm infringes a constraint, and can only be justified if enough good (or bad) outcomes are at stake.54 This is not to say that the harmed-person’s guilt is irrelevant. As we argue below, a person’s moral responsibility for creating the circumstances that call for infringing the constraint should plausibly affect the threshold function used to determine the permissibility of the infringement.55
Since the measures democracies employ in their fight against terrorism often infringe the deontological constraint against actively/intentionally harming people, one should assess their permissibility through Threshold Functions. The following sections discuss the key choices one must make in constructing such functions.
D. Constrained Economic Analysis of Intended Harm