THE RULE OF LAW AND THE LIMITS OF ANARCHY

Abstract

Anarchy is often contrasted with law, order, or security. But anarchist societies, by which I mean societies that lack a monopoly of coercive force, need not be lawless. They can develop sophisticated legal systems that regulate the behavior of their members and protect their rights. International law, market anarchism, and other models of anarchism such as the one proposed by Chandran Kukathas already exhibit or could plausibly exhibit complex legal rules and institutions. I will show that insofar as these models rely on consent, they all share similar structural flaws, namely, that they cannot meet basic rule-of-law values such as equality before the law and access to legal remedies for wrongs that embody and respect individual moral equality, even minimally conceived. The implication of this argument is not to vindicate state-based legal systems. Rather it is to show that legal systems, state-based or not, must have a strong nonconsensual, coercive element: the process of making, applying, and enforcing law must, to some extent, be severed from consent if law is to perform its function of providing for minimal justice.

Anarchy is often contrasted with law, order, or security. Most social contract theorists justify the state as an alternative to a state of nature in which individuals enjoy little or no protection from law.¹ But anarchist societies, by which I mean societies that lack a monopoly of coercive force, need not be lawless. They can develop sophisticated legal systems that regulate the behavior of their members and protect their rights. They can have first-order rules that set limits on the way in which individuals can inconvenience and harm each other, and second-order rules about how to make, change, interpret, and apply first-order rules.² They can also develop sophisticated institutional machinery, such as courts and legislative assemblies, whose role is to administer and enforce the rules.

International law is an example of a consensual legal system that has developed in an anarchic world. In the absence of a global leviathan, states and nonstate agents have created legal rules from the ground up via multilateral and bilateral treaties that mainly apply with state consent. International law relies on multiple regional and international courts to settle disputes, and assemblies of state parties can make and change rules.³ International law does contain bodies of law that are said to apply without state consent, such as customary law, the rules of the UN Charter, and jus cogens norms, but by and large, consent plays a critical role in its operation. The Vienna Convention on the Law of Treaties (VCLT), a body of secondary rules that regulate the making, validity, and termination of treaties, states in its Article 34 that “[a] treaty does not create either obligations or rights for a third State without its consent.”

There are other possible models of consensual anarchist legal systems. Self-governing communities bound together by close cultural or religious ties can create legal rules around their core values.⁴ Still other anarchist models are based on law enforcement provided by competing private security companies.⁵ These various imagined visions of a world without a state can be described as free associationism, anarcho-capitalism, or variations of the two. I will show that they all share similar structural flaws, namely, that they cannot meet basic rule-of-law values such as equality before the law and access to legal remedies for wrongs, and that to the extent that they do, they cease to function as anarchist legal systems. In consensual legal systems, consent (or lack thereof) can decouple the authority of the legal rules from that of the institutions that apply and enforce them. Legal subjects can refuse to give consent to adjudication and enforcement institutions even when they consent to rules, resulting in the uneven application of the latter and unequal accountability for rule-breaking. The case of international law shows how acute this problem can be. I will not defend these rule-of-law features here, but rather will assume their centrality in a minimally just, working legal order.

The implication of this argument is not to vindicate state-based legal systems. Rather it is to show that legal systems, state-based or not, must have a strong nonconsensual, coercive element that places rule-of-law features at their core. The process of making, applying, and enforcing law must, to some extent, be severed from consent if law is to perform its function of providing for minimal justice. Compared to the state-based legal systems that offer more robust rule-of-law protections, namely, liberal democratic systems that emphasize equality of legal rights and the compulsory jurisdiction of the institutions that interpret and apply them, consensual legal systems cannot enjoy the uniformity and consistency necessary to embody and respect individual moral equality, even minimally conceived. I leave open the possibility that anarchist systems organized around nonconsensual principles may be able to overcome these challenges.

This finding invites us to reevaluate our concept of law and consider more carefully which features are necessary for functioning, just legal orders. It provides a new justification for a series of features often considered necessary, but whose necessity is often misunderstood, namely, the uniform protection of legal subjects—and therefore a new justification for the generality of rules and comprehensive legal jurisdiction. The structural failings of consensual anarchist legal systems show that the justification of these features rests on essential demands of legal justice.

I. LAW UNDER INTERNATIONAL ANARCHY

Central to anarchist thought is the idea that state authority is in tension with individual autonomy. There are various reasons for this, including the absence of actual consent of most individuals to the authority of the states they happen to live under, the belief that states as institutional structures mostly serve the interests of capitalist producers, and the belief that state services are not subjected to the discipline of the market and are thus inefficient and wasteful. Whatever their reasons for rejecting state authority, anarchists share the view that there is a deep tension between individual autonomy and state authority. Robert Paul Wolff has succinctly captured this idea in his well-known essay In Defense of Anarchism:

The defining mark of the state is authority, the right to rule. The primary obligation of man is autonomy, the refusal to be ruled. It would seem, then, that there can be no resolution of the conflict between the autonomy of the individual and the putative authority of the state. Insofar as a man fullfils his obligation to make himself the author of his decisions, he will resist the state's claim to have authority over him. That is to say, he will deny that he has a duty to obey the laws of the state simply because they are the laws. In that sense, it would seem that anarchism is the only political doctrine consistent with the virtue of autonomy.⁶

Like many other anarchists, Wolff believes that individuals cannot simultaneously retain authority over their decisions and accept the authority of the state. Wolff does not merely see a tension between autonomy and state authority, which could be solvable in practice, but rather a fundamental incompatibility, which makes states inimical to individual autonomy as such. This is, in a nutshell, the anarchist's position on the importance of individual consent to rule.⁷ I think this position is wrong, but I will not challenge it directly here. Rather I aim to explore the consequences of the requirement of individual consent for the authority of law and demonstrate some of the serious problems it raises. The argument will provide indirect evidence of the idea that autonomy can only be exercised in the context of a developed legal system, chosen or unchosen, with certain rule-of-law features, which protects one's decision-making authority by creating reciprocal constraints on behavior. But the focus will be on showcasing the rule-of-law limitations of consensual anarchist legal systems.

The rule of law captures formal and substantive ideals in a legal system, or what has been termed the “internal morality of law.”⁸ Rule-of-law ideals constrain the arbitrary use of legal and political power for the sake of individual freedom and equality. Among the most important principles of the rule of law are (1) generality, (2) publicity, (3) prospectivity, (4) stability, (5) capacity to reflect clear, reasonable, and mutually consistent demands on individuals, (6) proportional punishment, (7) easy access to courts, and (8) an independent judiciary. For reasons of space, I will only focus on two related principles in this article, namely, (1) generality, which requires that laws be applied equally and that no one is above the law, and (7) easy access to courts, which requires that resort to justice is accessible to those who have been wronged through a court system designed to facilitate speedy and equitable resolution of complaints. These two principles depend on and reinforce each other: the fact that some members of a community are not able to access the justice system for redressing wrongs committed against them undermines the “equality before the law” requirement, and the fact that the law does not apply uniformly and generally often translates into inequitable access to courts. They also represent the criteria whose absence signals most clearly the moral failings of consensual legal systems.

In this section and the next, I discuss the rule-of-law limits under two anarchist systems: international law and market anarchism, also known as anarcho-capitalism, both of which rely on strict notions of consent for the authority and applicability of law. I will show that law under these anarchist systems shares the following structural features: legal subjects do not enjoy equality under the law, and they are unable to hold perpetrators accountable for violations. While in most legal systems these features of law are realized in degrees across the whole range of legal rules, in a well-developed legal system the presumption of equal access to remedies for wrongs is strong, because it is built into the design of rules and legal institutions. I will show that under anarchism these features fail to obtain consistently even in instances affecting very basic rules such as the security of persons. These features are replicated in other anarchist systems that relax to some extent the requirement for individual consent.

One of the distinguishing features of international law compared to state-based legal systems is that, for the most part, legal subjects, namely, states, must give consent to both the rules and the institutions that interpret and apply those rules. This is seen at once as normatively desirable, since a consensual process protects state sovereignty, and inevitable, since there is no centralized law-making and law enforcement. In principle and in practice, this means that contractual processes through which legal subjects enact legal rules can divorce consent to rule-making from consent to adjudicating and enforcing institutions. States can make rules but refuse to empower courts to interpret them and settle disputes or to empower enforcement mechanisms that ensure their compliance.⁹ For example, some treaties simply lack interpretive and dispute resolution mechanisms. Others include dispute resolution mechanisms, but through the practice of reservation states can reject these mechanisms’ authority to make decisions without their consent even when they agree to the rules set down by the treaty. Without interpretive and adjudicative institutions, legal subjects can resort to conflicting interpretations of rules and can refuse accountability for violations because they can refuse the jurisdiction of a court that can settle disagreement.

In saying that states can refuse accountability for violations I do not mean to suggest that they are always successful. Indeed, the scholarly literature documents some of the ways in which states are held responsible for wrongdoing. Oona Hathaway and Scott Shapiro show how states that are perceived to be breaking the rules are subject to “outcasting” by other states, which consists in the withdrawal of benefits from international cooperation and some institutional schemes through sanctions, boycotts, severing of diplomatic relations, and countermeasures.¹⁰ Andrew Guzman discusses the mechanisms that lead states to respond to and internalize the rules of international law and that make sanctioning mechanisms by other states effective through reputation, reciprocity, and retaliation.¹¹ These means of enforcement of international rules are decentralized in the sense that they rely on states to individually or collectively provide incentives to comply and impose costs on noncomplying states. As Liam Murphy and others observe, the fact that the sanctioning system of international law is decentralized should not impugn its lawlike credentials.¹² But there are of course many cases in which states are not willing to bear the costs of sanctioning noncompliance, either because they do not have the capacity or they cannot afford to expend the necessary resources. Gregory Shaffer documents the disproportionate costs that developing states must bear when bringing cases before the World Trade Organization (WTO) dispute resolution mechanisms, evidenced by the fact that “95 of the WTO's 120 non-OECD members had never filed a complaint before the WTO,” and by the fact that close to half of all the filings come from the United States and the European Union.¹³ It is most unlikely that the states that have not filed complaints have no serious grievances to settle via the dispute resolution mechanism. The more likely explanation is that they lack both the legal expertise and the financial resources needed to push a claim through and persist in the face of pushback from wealthier states during what is often a lengthy, drawn-out process. This means that developing countries have a much harder time protecting their legal rights in international law, because their ability to sanction more economically or militarily powerful states is severely limited by their ability to bring them before a dispute resolution mechanism or to outcast.

This is clearly the case even when a decision of a court or arbitral tribunal unequivocally establishes the fact of a violation, as is the case with the South China Sea. China is making excessive claims to extend its exclusive economic zone while Vietnam, the Philippines, Malaysia, and Indonesia are unable to effectively push back against these claims.¹⁴ A more significant problem with a decentralized sanctioning mechanism is that, in cases in which courts lack the relevant jurisdiction to make a determination of fact regarding violations, forms of accountability such as outcasting rely on private judgments about the noncompliance of other states, which is a form of vigilante justice, rather than on legal findings in which due process establishes public grounds for outcasting.

The difficulties with vigilante justice are manifold, and they relate to the fact that private agents use their own judgments to enforce compliance. Vigilante justice is not only wrong, or not primarily wrong, because those who go after rule-breakers engage in unauthorized usurpation of public power. Rather, private judgments could be: (1) wrong on the facts, (2) biased/self-serving, or (3) illegitimate in the eyes of others and subject to retaliation/challenge. In practice, these problems are attenuated when the enforcement action is multilateral. But multilateral action can still be wrong on the facts and biased. The second Iraqi war arguably suffered from these problems.¹⁵ This does not mean that institutionalized nonconsensual enforcement is flawless, but that private action to enforce the law can predictably have these undesirable features. This is a point that is open to conceptual and empirical challenge, no doubt. But it is a point with deep Lockean roots that is a neglected insight in the anarchist literature, whether international law–related or not.

International law has developed in an anarchic world without a centralized power to make rules and enforce them. Thus, by necessity, it is created by and applies primarily with state consent, and according to the rules of the VCLT, the convention regulating the making, terminating, and validity of international treaties, the consent has to be explicit (see, e.g., Articles 6–41). States choose which treaties to agree to, and moreover choose which parts of a treaty apply to them through the practice of reservations. For example, the provision of the Genocide Convention with the most reservations is the one giving the International Court of Justice (ICJ) jurisdiction over interpreting the rules of the Convention, ascertaining the fact of a violation, and deciding what if any remedies are imposed on violators.¹⁶ The same is true of many other international treaties in which states withhold consent from provisions enabling dispute resolution and interpretation of the rules, including, ironically, the VCLT itself.¹⁷ Article 66 allows a state party to initiate a case before the International Court of Justice in situations in which a treaty is thought to conflict with a peremptory norm that makes it invalid. This article gives states the capacity to submit disputes regarding certain provisions of the treaty to the ICJ jurisdiction without the consent of the other party, just as individuals can under domestic law. The numerous reservations to it deny that such unilateral legal action is justified. For example, Cuba's reservation reads:

The Government of the Republic of Cuba enters an explicit reservation to the procedure established under article 66 of the Convention, since it believes that any dispute should be settled by any means adopted by agreement between the parties to the dispute; the Republic of Cuba therefore cannot accept solutions which provide means for one of the parties, without the consent of the other to submit the dispute to procedures for judicial settlement, arbitration and conciliation.

Cuba is in effect resisting the jurisdiction of the ICJ in matters related to certain provisions of the VCLT without its consent. Similar reservations have been registered by Algeria, Armenia, China, Russia, Saudi Arabia, Tunisia, and Vietnam, among others.

The importance of this insistence on consent for the jurisdiction of the ICJ for interpretive issues related to the rules of the VCLT is that differing interpretations of the rules will persist and serious violations will remain difficult to establish in the absence of the noncomplying state's consent. The overemphasis on consent means that, as Miodrag A. Jovanović puts it, since “there is no compulsory system of adjudication, . . . adjudication is the exception, not the norm.”¹⁸ The fact that “the ICJ is only competent to entertain disputes between states that have accepted its jurisdiction,” means that, for the most part, “the ICJ is arguably serving as ‘a glorified arbitration panel.’”¹⁹ This is true of other courts and tribunals as well, despite the significant rise in recent decades in their number, and in the increasingly complex roles they adopt as interpretive, administrative, and enforcement mechanisms.²⁰ States in effect have veto power over the authority of international tribunals empowered to administer legal justice. An instance of this power is exemplified by states that withdraw consent to international judicial organs in the wake of unfavorable decisions, as the United States has done with respect to the International Court of Justice in the wake of the Nicaragua and Avena decisions.²¹

International criminal law is emblematic of this veto power embedded in the consensual character of international law. Treaties that establish individual criminal accountability for crimes of concern to the international community as a whole, including genocide, crimes against humanity, and war crimes, only apply with states’ express consent. Consent is usually provided through ratification according to national legal procedures. The International Criminal Court (ICC) established via the Rome Statute (1998), which came into effect in 2002 after a sufficient number of states ratified it, only exercises authority over member states. Countries such as the United States, China, Russia, and Sudan are not members. The only exception to the consent-based jurisdiction of the ICC is Security Council referral, which is how the ICC came to issue an arrest warrant for the then-president of Sudan Omar al-Bashir for crimes against humanity and genocide.²² Since China, Russia, and the United States are permanent members of the Security Council with unilateral veto power (again, a strong consensual element), the latter is unlikely to issue any calls for prosecution of the permanent members or their allies. Indeed, China and Russia have consistently vetoed efforts to reduce the capacity of the Syrian government to inflict harm on its own citizens via the creation of no-fly zones and other humanitarian efforts, due to their political and economic interests in the region.²³

Ad hoc international criminal tribunals like the International Criminal Tribunal for the Former Yugoslavia or International Criminal Tribunal for Rwanda are usually established by the United Nations Security Council without the consent of the state or states that fall under their jurisdiction. These tribunals have delivered an important nonconsensual measure of justice, but they are designated “ad hoc” for a reason: they are established for a specific conflict, with limited temporal and geographical authority, and they cease to exist after a number of years. The establishment of such tribunals is selective, and many humanitarian emergencies marred by large-scale violations of human rights lack any such accountability mechanisms.

The effect of this consensual element of international criminal law is that the representatives of states that do not consent to the rules of international criminal law or to the authority of the legal institutions that interpret these rules cannot be brought to justice, leaving the victims of these very serious crimes without the possibility of remedying the wrongs against them. This is not entirely surprising: the rules of international criminal law usually hold individual public officials accountable for serious crimes, yet the jurisdiction of institutions of accountability depends on some of the same individuals giving their consent in the name of the country they represent. We can see this tension at work in the withdrawal of the Philippines and Burundi from the membership of the ICC just as the latter was opening an investigation into president Duterte's extrajudicial killings and had been investigating officials in Burundi for crimes against humanity.²⁴ State officials may have other, principled reasons to object to the authority of the ICC.²⁵ However, a system of criminal law relying almost exclusively on state consent creates too many veto points that protect the individuals and institutions in a position to inflict grievous harm, and leaves people worldwide exposed to threats to life and security with little possibility of redress.²⁶

One may wonder whether the ability to leave international treaties at will and thus undermine the jurisdiction of international institutions such as courts over particular states at will is a bug rather than a feature of a consent-based legal system. Surely institutions can be designed such that withdrawal is more difficult, takes longer, or can only take place under certain conditions. This is of course true, but these design features raise states’ costs of agreeing to treaties in the first place, which makes the likelihood of them joining lower. Increasing exit costs simply moves the problem of consent to a different level. At whatever stage consent is granted, it renders fragile and unstable the legal subjects’ (in this case states’) relationship with the legal institutions that ensure that rule-of-law values are upheld.

International law is instructive as an anarchic legal system because it shows the ways in which consent can undermine the rule of law: consent to rules can be divorced from consent to interpretive and adjudication mechanisms, thus insulating legal subjects from accountability for wrongs, even when the rules are widely adopted. No criminal law system can operate by relying almost exclusively on the consent of the legal subjects. This is because legal subjects with a tendency to disregard or break the rules of criminal law will refuse consent to either the rules or the institutions that enforce them.

The most important implication of this basic point for international law is that there is an important trade-off between respect for state consent as an embodiment of respect for state sovereignty and the rule of international law. It may be the case that state sovereignty is incompatible with the authority of coercive international institutions, just as Wolff claims is the case in the conflict between individual autonomy and the authority of the state. Nonetheless, if we decide to privilege state sovereignty, we ought to be aware of its cost in terms of the rule of law. States bound to international law via consent make the equality of states before international law unattainable.

Some regional courts do not privilege state sovereignty, although they give a margin of appreciation to states regarding how decisions are implemented. However, it is one thing to leave a margin of appreciation that protects state sovereignty, as the Court of Justice of the European Union or the European Court of Human Rights does, when their jurisdiction is compulsory (for member states only) and states do not have the freedom of opting out of disputes. It is another thing to say that the margin of appreciation or the doctrine of subsidiarity advocates in favor of international courts such as the International Court of Justice or the International Criminal Court being denied compulsory jurisdiction. The latter is a position that defends state autonomy at the expense of the rule of law.

The design of international law will be structurally similar to the design of other consensual legal systems, and, therefore, we can expect the latter to have the following features: rules, even basic rules regarding the protection of basic rights, will not apply equally to all, since potential subjects are free to withhold consent to them; and, relatedly, when violations of basic rules protecting human life and security occur, victims will be unable to bring perpetrators to justice when the latter do not agree to authority of the institutional mechanisms for the interpretation and enforcement of those rules. I will showcase the rule-of-law problems of market anarchism next.

II. VIOLENCE AND LEGAL REDRESS UNDER MARKET ANARCHISM

To see why these failings would arise under market anarchism as well, let us consider this model in more detail. Like most anarchists, market anarchists reject state authority for a combination of reasons. Chief among them is the lack of actual consent to state authority. In addition, they share the concern that in the absence of market discipline, the allocation of goods and services by the state for individual security and protection, and even the production of law, will be arbitrary, i.e., lack any connection to the actual needs of the individuals for these services and to a proper concern for relative costs of production.²⁷ Markets adapt dynamically via the price mechanism to align supply and demand, namely, to match individual needs with the requisite volume of goods and services. Over time and through a process of learning from failure and competitive pressures, producers select the most economically efficient combination of inputs to meet the demand for outputs. Without the competitive pressure of markets and prices, states have no idea what the most effective mechanism for provision is or how much law or security is needed, and will undersupply or oversupply these goods, vastly misallocating resources in the process.

A solution that both preserves individual autonomy and ensures that individuals receive the type of service that best corresponds to their needs involves private security companies. In a world of states, each state commits to more or less the same level of security for all individuals living on its territory. By contrast, in the absence of states, private security companies will customize the amount of security required by each individual according to her needs, and, crucially, her ability to pay.²⁸ Individuals will contract protection services just as they contract the services of an insurance agency or a moving company. And over time the market in protection services will diversify to cover a range of security needs and abilities to pay. In fact, market anarchists envision that not just security but all other government services, including all public goods, will be provided via market-based competition.

This radical approach to individual freedom raises a whole host of questions about accessibility for those of low economic means, human welfare, the nature of communities likely to exist under this system, and rights protection. Although market anarchists have better and worse answers to some of the possible challenges to their model of society, I will set most of them aside to focus on the realization of rule-of-law principles under a market anarchist system. The first thing to note is that market anarchists imagine law production to be taken over by private security companies as a complement to or in addition to their security functions, so individuals will contract into different legal systems by virtue of contracting into different private security companies.²⁹ The second thing to note is that protection services will likely operate on a nonterritorial basis just as markets in most goods and services do. This means that neighbors could acquire the services of different protection agencies and thus would place themselves under the authority of different legal systems.

Although a world of private protection agencies may seem distant from the world of international law made by states, in practice the contractual nature of law creates similar structural problems in both systems, namely, the separation of rule-making from rule adjudication and enforcement. Anarchists imagine that private protection agencies will create rules for their members as a condition of service and that these rules in effect will constitute law for them.³⁰ Since rules will be necessarily general and abstract, private security companies will have to find ways to provide mechanisms for making and interpreting the rules, determine whether specific behaviors fall within the purview of a given rule, and decide on remedies for law-breaking.³¹ And they will have to find ways to carve jurisdictional authority among them.

But individuals need not acquire all of the law-making and law-adjudicating services a private company offers. Private protection agencies will offer a differentiated range of services, from those that offer some minimal, round-the-clock physical protection, as many do now in parallel to state provision, to those that combine legislative, judicial, and enforcement services. They can also choose to subcontract some of these services to other companies. Crucially, all of the relationships among individuals and various providers of legal services can be fully traced to the various contracts individuals and agencies in the system enter into with one another. But the fact that individuals are tied to these private legal systems via a strong consensual system of market contracts raises special problems for rule-of-law ideals under anarchy. We will consider two distinct cases: when individuals are members of the same private security company, and when they are not (either they are members of different private security companies, or one is a member of one and the other is a member of none).³²

The first scenario concerns an accusation of serious violence to person or property between two individuals who are members of the same private security company. The first thing to note is that, even under anarchy, some individuals can enjoy strong rule-of-law protection. They will be able to receive justice for the wrongs dome to them and enforce a ruling of harm and compensation as long as they can afford comprehensive legal services on a contractual basis and the perpetrator is a member of the same company. Private security companies could conceivably provide a whole range of complex legal services, from rule-making to interpretation, dispute resolution, and enforcement, and could adopt the policy that to be a member one must adopt the full package of legal services and be subject to that company's compulsory jurisdiction. Recent research has shown that private, decentralized legal systems can provide public order in much the same way that centralized systems do. Gillian Hadfield and Barry Weingast, and many others, have discussed the conditions under which decentralized law without the state is possible and how equilibrium can emerge in the absences of states.³³ The question therefore is not whether security and order are possible under anarchism, but whether their administration through the legal system is determined in accordance with rule-of-law values of equality under the law and legal access to remedies for wrongs.

This is not likely to be the case when individuals choose private security companies with less restrictive policies and will be free to subscribe to the protection agencies that will offer differentiated packages of services, some of which will not involve compulsory jurisdiction to settle disputes among members. Providing rules, interpreting them, and enforcing them are distinct, specialized services, and companies could conceivably bundle these services differently. Market anarchists may assume that individuals will acquire all of these services in a bundle, since it will make more economic sense for individuals to subscribe to all,³⁴ but it is not clear that this is the case. Some companies may offer these services only as a complete package, but it is imaginable that others will be able to make money only if they offer basic security services. This is because individuals will vary in how much they plan to rely on self-help in their dealing with other individuals, and they may prefer to do so in some areas of law and not others. Since the services of private security companies are demand driven, how these companies bundle legal services will depend in part on what services individuals in a position to acquire them prefer to purchase.

Therefore, a first hurdle faced by an individual is to bring a case before a court or arbitration tribunal and make a determination of rule violation. In the case of individuals who are covered by the same set of legal services and institutions, the process can operate smoothly and even justly as long as complex legal services are bundled and compulsory jurisdiction is accepted, and the company itself creates fair, impartial rules and abides by due process and rule-of-law principles. Thus, anarchy is compatible with extensive rights protection and respect for rule-of-law principles, at least for some individuals. Nonetheless, we can imagine a situation in which an individual who receives an unfavorable judgment decides to withdraw consent mid-trial, or after a judgment has been entered against her but not enforced. Under market anarchism, legal institutions are just as fragile and vulnerable to the vagaries of individual consent as they are in international law. Private security companies will have rules about the conditions in which individuals may end the contractual agreement for services. But contracting parties will disagree about when these conditions are met, and they will therefore need to resort to third parties for arbitration and enforcement. In most legal systems, arbitration is divorced from enforcement; therefore, individuals and private security companies will have to agree ex ante to the authority of a third-party enforcer of their contract, likely via a separate contract. Violations of that contract also need third-party adjudication and enforcement, and so on.

Unless this regress ad infinitum is resolved, the wronged individual will not be able to rely on accountability for a violation of a legal rule. They can attempt to bring a case before an arbitration tribunal, but there will be no guarantee that the defendant is bound by the same rules and jurisdiction, or that they will continue to consent to them even if these are shared. One objection to this critique will be that private security companies will not be able to effectively protect their members unless they adopt some mechanism for compulsory adjudication and enforcement, and thus resolve the problem of infinite regress. But this cannot be true. Private protection companies will offer a variety of services. It is conceivable that some individuals will contract to buy some but not others. In that case, imposing a compulsory arbitration mechanism on individuals who have not consented to it will amount to a serious violation of individual autonomy from an anarchist standpoint. Indeed, the whole purpose of a system of private contracts by which individuals organize their lives is to avoid unwanted, unconsented-to imposition of legal authority.

What this means is that the contractual nature of individuals’ relationship with law in its various dimensions means that individuals could be conceivably bound by different laws, and when they are bound by the same law, they can opt out of judicial and enforcement mechanisms for those laws. This is the case in international law, where states’ bundles of legal rights and obligations depend on the treaties they consent to, and even two states bound by the same treaties may differ in their acceptance of the authority of judicial and enforcement mechanisms for it. Just like in the case of international law, under market anarchism individuals who had been seriously harmed could not hold perpetrators accountable for their actions and would thus be left vulnerable to abuse. But the fact that individuals will be subscribers of different protection agencies will mean that even the assumption of shared rules is too strong.

In the case of purported violence between the members of different protection agencies, market anarchists assume that either different protection companies will converge on rules, or conflicts over rule violations will be resolved through the willingness of the agencies in question to submit themselves to common arbitration and thus resolve disagreements between them and their members peacefully rather than by violence.³⁵ Neither of these assumptions is defensible. First, market uniformity understood as standardization does indeed happen in some markets (the physical size of the credit cards is the same regardless of what bank issues them), but markets are characterized just as much by differentiation as they are by standardization. Indeed, what makes markets appealing as mechanisms for the delivery of security and legal services is precisely their ability to differentiate their services to match the varied needs of the consumers, and systems of rules and the institutions needed for their enforcement will be no different. There is no reason to think that various private protection agencies will converge on the same set of legal rules, and in fact variance across existing legal systems shows that the opposite is likely to be the case. Second, the willingness to solve disagreements peacefully is assumed as a consequence of the incentive these companies have to keep costs down. And this incentive is demand driven: individuals want to economize on private security.³⁶ In a world of conflict among private security companies, their costs will be high, and as a consequence the premiums for individuals will be high.³⁷ But in the real world, individuals have multiple incentives that interact with each other in complex ways. While they may prefer cheaper security services over more expensive ones, they may also have a stronger preference to overpower, take over, and even exterminate other religious, racial, and ethnic groups.

Individuals with such malign, belligerent preferences can choose to pay a premium to private security companies that are willing to act on them, and thus companies can survive and thrive in an environment where peaceful resolution of conflicts is not valued and in fact is actively discouraged. In this kind of world, individuals who are wronged will again be left with nowhere to turn to redress injustices committed against them. Note that the issue is not whether private security companies could conceivably agree to common arbitration and enforcement. Even granting that they could, their propensity to do so will be contingent. It will depend on a large number of factors, including their members’ preferences. Since these preferences can be malign, and we can assume individuals will be willing to devote large resources to satisfying them, private security companies under market anarchism can coexist in violent environments. Under these conditions, individuals may lack access to mechanisms for the administration of basic criminal justice.

This means that just as under international law, basic equality before law does not obtain under market anarchism, except accidentally. Wronged individuals have little recourse for violations of their rights, unless they and the perpetrators are full members of the same complex legal order characterized by strong, nonconsensual interpretive and enforcement mechanisms. For market anarchism, this means that individuals can hope to have access to justice only when disputes arise between them and other members of the same private security company that also offers complex legal adjudication and enforcement for members, and for states only when they are members of the same treaties with robust judicial mechanisms. And importantly, these rule-of-law features only obtain as long as consent by all parties to the same security company or international treaty is upheld. But they will not obtain with the same regularity for individuals who are members of different security companies, or states that are members of competing treaties.

Market anarchists will object that there can be no parallel to international law since the treaties and institutions of the latter are not based in the decisions of market firms operating in a competitive market. This is an important difference, but I do not think it affects the relevance of the comparison that I draw. The fact that one is a system based on market competition and one is not will affect the type and range of institutional options available in each system, but not how individual legal subjects are bound by law in them. This last element depends primarily on consent, and in both systems legal subjects are tied to law and law enforcement on a contractual basis. Indeed, the VCLT deems states entering international treaties “contracting parties” (see Article 2) just as individuals would be under a system of completely privatized legal services. In practice this means that legal subjects can withhold consent from institutions that set rules of behavior and hold them accountable for wrongs inflicted on others. And in both systems, consent allows them to escape accountability for wrongs.

We will turn next to an additional model of anarchism that relaxes the requirements of individual consent. I will show that in this system law suffers from similar structural problems, and that to the extent that it does not, the system is in fact no longer anarchist. I do not intend to present this as an analytical truth. Rather, I will show that insofar as anarchist legal systems protect equality before the law and access to remedies for wrongs, they develop legal authorities and legal hierarchies that begin to resemble the nonconsensual legal systems of modern states.

III. THE RULE OF LAW UNDER FREE ASSOCIATIONISM

Other anarchist models emphasize social life organized around small self-governing communities, freely entered into and in which individuals exercise robust rights of exit. The relationship among communities will have features similar to international law, to the extent that intercommunity rules will be binding only on the basis of individual communities’ consent that can be withdrawn at any time. But individuals can escape accountability in their communities if a strong form of individual consent is the basis of the community's legal jurisdiction over its members. As long as individuals can opt out of their community's rules, they can escape legal accountability. I will show these structural similarities for a religious/cultural form of anarchism organized around shared cultural or religious values in a model that Chandran Kukathas calls the liberal archipelago. A liberal archipelago consists of communities free to choose diverse modes of political organization and in which no one form of political authority will take priority over any other. What we have instead is a community of communities with overlapping jurisdictions.³⁸

Kukathas puts forward a distinctive liberal vision of a free society, marked by pluralism of religious and cultural communities. The anarchist aspect of these communities lies in the freedom of their members to associate and disassociate from one another to pursue common visions of politics and social cooperation; thus I will call these anarchist visions free associationism. Individual consent is mainly expressed through the right of exit, but the conditions of individual consent are relaxed since the main decision-making unit is the community, not the individual. Individual consent is still important, but it is not expressed through a contract and is instead presumed in the absence of exit. Consent to the group reflects the unity of shared social and political vision within each community.

Liberals believe a political regime that protects individuals as free and equal will reflect and foster the diversity of conceptions of the good life that inevitably arises in society. But Kukathas goes further than most liberals and argues that this diversity will lead to variability in political arrangements as well. Individuals exercising their rights of association and dissociation will form communities whose institutional arrangements and political organization vary and change rather than “uphold[ing] a determinate set of institutions within a closed order.”³⁹ What makes the resulting conglomeration of associations an anarchist order is the fact that associations formed according to individual people's freedom to associate with each other will form “an archipelago of competing and overlapping jurisdictions”⁴⁰ and will be constrained by a principle of mutual toleration. Political association is one among many other associations, and there is no hierarchically superior place it occupies over diverse religious and cultural communities. The metaphor of an archipelago does important work here. Communities are like islands that show deference to each other's internal political system and external boundaries. In the archipelago “there will be a multiplicity of authorities, each independent of the others, and sustained by the acquiescence of its subjects.”⁴¹

The free society of the archipelago requires the freedom to move between communities, which suggests that a central mechanism that enables the functioning of the archipelago is individual consent, or rather “acquiescence,” as Kukathas puts it. Acquiescence is not as strong as consent, or at least not as strong as express, contractual consent. Acquiescence involves a more passive stance of not actively resisting one's authority and can be inferred from the tacit acceptance of one's political order. A lack of consent is signaled by individuals choosing to exit.

In Kukathas's model, there is an intentional elision between the cultural and political dimension of a community. Cultural and associational diversity constitutes the basis of the archipelago and of the practice of overlapping jurisdictions. Yet there is considerable difficulty in translating the metaphor of the archipelago for the world of politics, because it becomes much harder to understand how to distinguish between these communities’ various functions. For example, it is not clear what counts as a political or legal institution as opposed to say a cultural or religious one in any community that is part of the archipelago. The idea is not to suggest that they must always be differentiated, but that we need to understand better how cultural and religious diversity translates into diversity of legal jurisdiction, if it does. It is certainly possible, as liberal democratic states show, for diverse cultural and religious communities to share the same legal system. Focusing on the nature of law and legal authority in this liberal archipelago, and the relationship of individuals and communities to law, may give us a better sense of how to understand these distinct communal functions.

The first indication of the legal order of the archipelago comes from Kukathas's description of an order of competing jurisdictions. The archipelago is “a form of order in which authorities function under laws which are themselves beyond the reach of any singular power.”⁴² While there are various legal systems that may overlap, their interaction takes place via a form of horizontal equality and respect for one another's boundaries without interference or control by one over the others, and this includes the legal order of the archipelago itself. Thus, in the archipelago there is no authoritative law that is hierarchically superior (as it would be in a federation) to the laws of the various islands in the system. Or rather, while the archipelago may have some rules, and they overlap with the rules of the component communities, it does not have superior status to regulate those communities from the outside.

However, it seems that at times Kukathas admits to the importance of some common rules that take priority: the archipelago is “a collection of communities (and, so, authorities) associated under laws which recognize the freedom of individuals to associate as, and with whom, they wish.”⁴³ These laws recognizing freedom of association would presumably acquire hierarchically superior status to other local laws. Indeed, there would barely be anything liberal about the archipelago otherwise. But this means that even the limited liberal order that Kukathas envisions will be characterized by legal hierarchy. To ensure freedom of association, archipelago-level institutions must develop at least some minimal rule-making capacity, courts with jurisdiction to interpret and apply rules of free association and sanction breaches, and enforcement mechanisms to restore and protect the legal rights so generated, including rights of exit.

Moreover, such authority cannot simply be limited to institutionalizing the protection of freedom of association for individuals. It will have to concern itself, inevitably, with regulating the relationships among communities. As with all other forms of anarchism, we cannot assume that communities will observe self-restraint in their interactions with one another. There may be cases in which one or some communities will seek to usurp others’ legal prerogatives and boundaries. While mutual toleration between communities is a corollary of freedom of association, it is not a self-enforcing principle. The central legal authority will thus have to create and enforce rules relating to the use of force and the principle of noninterference, among others.

Legal hierarchy based on the authority of archipelago-wide institutions will generate legal uniformity and unity among legal rules protecting basic rights of freedom of association and exit and managing the interactions of communities embodying those rights. And to be effective, those legal rules will have to have independent power and authority over the component units and individuals as subjects of those legal rules. The only way to ensure the existence of an archipelago is to grant those legal rules supremacy over the rules of the component units, much like federal law has supremacy over state law in federal systems. Supremacy is required for the archipelago to exist at all.

Thus, contrary to Kukathas, there is a singular legal system that provides the context for and has priority over all other legal systems. That legal system assures the rules of mutual interaction among the units in the system and preserves the freedom of association of individuals in it by ensuring rights of exit against their communities. This system describes, with some differences, the one that exists today, namely, that of international law. States are part of a legal system that restricts the use of force and protects their sovereign prerogatives against interference from other states. International law does much more of course. It regulates territorial boundaries, helps solve complex cooperation and coordination problems among states, and provides general protections for individuals in the form of human rights.

Not surprisingly, Kukathas notes the similarities between his anarchist vision and international politics. In his view, the metaphor of the archipelago describes the international community of states as a “society of societies” that lacks a singular authority regulating the activities of all of the component societies.⁴⁴

This is why Kukathas believes that the international society “is a kind of liberal society since it is a society of multiple authorities operating under a de facto regime of mutual toleration.”⁴⁵ He goes on to say, “What keeps this order liberal is the fact that there is no hierarchy under which the authority of the various states is subordinated; and the fact that the norm of free exit is dominant, since most states do not recognize the right of other states to keep their subjects within their borders against their will—even though many states do, in fact, assert such a right.”⁴⁶ The international society is the ultimate liberal archipelago. Yet while liberal in the narrow sense Kukathas advocates, the international society is deeply defective as a legal order in the ways we have just seen.

States can extricate themselves from accountability from wrongdoing, although outcasting and other reprisals will limit somewhat the extent to which states can enjoy absolute freedom in this regard. Today states by and large accept the prohibition against genocide, crimes against humanity, and war crimes. Despite this agreement in principle to limit the authority of states to harm their own citizens codified through various treaties and conventions that prohibit these crimes, little follows for noncomplying states. It is difficult to see how the United States, especially, would be constrained despite its acceptance of these international crimes, since it is not part of the ICC, it has the possibly to veto referrals to the Court as a permanent member of the Security Council, and it has withdrawn its blanket grant of authority to the ICJ, the only other international court that could potentially exercise authority over the United States in relation to these crimes. In this and in other areas of law, it is clear that states are not equal before the law, as some of the current members of the ICC correctly point out. And states that are actually bound by these norms through institutions of adjudication and enforcement are mostly bound due to their own consent. Similarly, in the liberal archipelago, much like in international law, it will be difficult to hold communities accountable for wrongdoing and even for violating the most fundamental legal and moral commitments of the archipelago.

Kukathas's free associationism will lack adequate protection for individual and group rights in the absence of a well-developed, institutionally differentiated legal system with the capacity to assert jurisdiction over disputes and demand compliance. These communities, along with the federations they form, will be deficient in basic rule-of-law values such as equality before law and access to remedies for wrongs. To the extent that they will develop coercive legal institutions with nonconsensual jurisdiction to prosecute and punish wrongs and make restitution for victims, they will cease to be anarchic and will come to resemble much more the hierarchical, nonconsensual structure of the legal systems of modern states.

But a deeper and more serious problem for a liberal such as Kukathas is that the liberal archipelago he envisions is compatible with significant levels of oppression and rights violations within the component societies, and oppression is the antithesis of the rule of law. This is due to two related features of the liberal archipelago: first, the only substantive right individuals enjoy in the liberal archipelago is the freedom of association, which in practice translates into a right of exit; and second, this right seems to lack effectiveness in the absence of institutions that articulate, protect, and restore the right in the face of violations.

Let's take the right of exit first. According to Kukathas, consent is preserved in this world, at least in principle, because individuals are “at liberty to reject the authority of one association in order to place themselves under the authority of another” and even “at liberty to reject the authority of one association in order to place themselves under the authority of another.”⁴⁷ Thus they can move between more liberal and illiberal societies. Liberal societies are those that tolerate internal dissent and multiple authorities, including legal authorities. Illiberal societies are those that do not.

Yet Kukathas explicitly rejects archipelago-level institutions whose role is to actively articulate, protect, and enforce this right against noncomplying communities. The only demand he makes is that the authority of communities to deny right of exit is not recognized by the other units.⁴⁸ A community can deny the right of exit to its members, as North Korea does, but other communities will enforce the restrictions against individuals who find themselves in the lucky position to be able to violate restrictions against exit. Kukathas advocates that communities behave more like South Korea than China. Restrictions on exit constitute a notorious tool of oppressive political regimes to maintain inhumane and degrading practices. It allows them to violate rights with impunity. Coupled with a norm of mutual toleration and noninterference into the affairs of other political associations, this means in effect that individuals are left at the mercy of their own political authorities with no recourse to outside institutions and no guarantees for their rights.

But even if the liberal archipelago would guarantee the right of exit, there would be considerable scope within it for oppression and rights violations. As Barbara Fried has shown, exit rights, with their corollary of tacit consent, would legitimate virtually all states, no matter how tyrannical.⁴⁹ Many of the worst human atrocities taking place today, including genocide, crimes against humanity, and war crimes, are happening in countries that do not deny their inhabitants the right to exit, including places like Myanmar, Yemen, Syria, China, and Sudan. Why is the right of exit not sufficient to protect individuals against abuses form their own communities? Individuals can have very close ties with their communities, which are difficult to break even in the face of widespread violation of their rights. Exit depends on their subjective evaluations about what level of injustice tips the balance in favor of leaving, and for many, exit imposes a huge personal cost. It is conceivable that some individuals never consider exit as part of their option set, either because they do not see themselves navigating new cultures successfully and taking on the tremendous economic risks associated with moving to a new country, because they have family in their care who are not mobile (very young, old, ill, disabled, etc.), or because they place too much value on their immediate community to imagine themselves being able to re-create it elsewhere. Those who do consider exit as an option might lack the resources or know-how to transition to a new social and economic environment, or they may find themselves unable to find communities willing to receive them. As Jacob Levy has recently put it, people “might have, literally, no place to go if they wish to exit the group into which they are born, no resources of space in which they could assemble their own dissident, hybrid, or rival association.”⁵⁰ Without a mandatory resettlement/refugee scheme imposed on all the component units of an archipelago, it is likely that many people who wish to exit their political community will be left without options, as the current plight of refugees aptly illustrates. And a mandatory resettlement/refugee policy that will force associations to accept people exiting other communities will likely violate the principle of freedom of association by forcing groups to associate with others they do not choose to associate with.

It is true that even today people exercise the right to exit from the most oppressive political regimes, sometimes en masse. But the majority of people in oppressive regimes do not exercise their right to exit even under the most trying circumstances. And when they do leave, their lives are not necessarily marked by sudden improvements. Exodus is perilous and even deadly, and it can be marked by indefinite transition in temporary camps set up to deal with large influxes of refugees. Lack of basic services and food, the prevalence of disease, and vulnerability to physical attack and sexual exploitations, or to the predatory schemes of human traffickers, are experiences common to those who exit oppressive regimes today. One cannot blame people from choosing the hell they know instead of the hell they can only imagine.

Therefore, the right of exit is consistent with significant and widespread oppression and rights violations. Tyranny and oppression, often marked by arbitrary rule and unequal rights, are the antithesis of the rule of law. Securing environments where individuals live violence-free and can flourish requires more robust and institutionalized systems of substantive rights, which enjoy institutional protection both at the level of each community and from the outside, at the level of the archipelago. Leaving individual communities responsible for the protection of their citizens means that when they fail, these citizens have no recourse to having their rights protected and cannot hold their communities accountable for wrongdoing. The consequence would be, as Adam Tebble aptly observes, to allow the exercise of “the most crude forms of cultural power” and give “carte blanche to all manner of cruel and offensive practices.”⁵¹

If the communities of the archipelago can be internally oppressive, the solution that recommends itself comes from the practice of existing liberal democratic societies, which institute legal systems that protect individual freedom and the rule of law.⁵² Interpreting and administering rules are specialized social functions that require personnel competent to decide on highly technical legal questions, thus likely requiring some form of specialized education and training, and who operate in institutions whose functions are well-defined and that themselves act under rule-of-law constraints. This means that in order to have a working, just system of law, communities will have to develop differentiated legal institutions, and those institutions will likely have coercive powers to act within their mandate. The presence of such institutions will create some form of hierarchy and possibly remove communities from the realm of pure anarchy and closer to the institutional apparatus of the modern state.

As Gregory Kavka eloquently showed, institutions of dispute resolution will be necessary even if all the members of a community share the same moral vision, namely, endorsing a set of fundamental moral values that underpin the organization of the community, as the members of the various “islands” in the archipelago are likely to.⁵³ This is because the moral vision will have to be hashed out in terms of specific rules about the boundaries of personal property, the protection of bodily integrity, and a whole host of political, social, and economic rights, and the same values and principles can be specified in a variety of ways in practice. Disagreements about how to specify, interpret, and implement them will exist even in a society of angels. Individuals who agree on fundamental moral values and principles will have different cognitive capacities and beliefs about the world and will draw different boundaries around certain basic moral values. They may disagree about whether smoking violates a right to life or whether abortion counts as killing.⁵⁴ Such disagreement will require rule-making that creates community-wide standards of justice and impartial dispute resolution mechanisms backed by coercive force in order to be effective.

But serious violations of the rule of law within communities may demand outside intervention. In an archipelago that is self-consciously pluralistic, individuals and communities will have different ideals of legal and political justice; only negotiated, publicly justifiable norms that protect the liberty and equality of all members can serve the role of a foundation for legal justice. But unless such norms are backed by specialized legal institutions ready to interpret them, determine violations, and impose remedies for victims, individuals and communities alike are left unprotected for serious wrongs done to them. The more the legal institutions acquire a permanent, standing authority backed by coercive force, the less they remain in the realm of horizontal, overlapping jurisdiction and the more they resemble the institutions of the modern state.

IV. CONCLUSION

I have made a broader case that anarchist politics is subject to dilemma: if law is meant to bind only voluntarily, on the basis of consent, then anarchist legal systems face basic rule-of-law problems. In the absence of consent to uniform rules against the most common forms of violence and wrongdoing, and to the jurisdiction of legal authorities with the power to interpret and enforce them, many individuals subject to violations do not have access to legal justice. To provide justice would require imposing on wrongdoers laws and institutions whose jurisdiction they have not agreed to, thus violating a core requirement of consensual anarchism, namely, that law only applies with individual consent. Yet to create the conditions of rule of law means departing from the voluntary basis of legal authority. Effective rule of law requires that nonconsensual legal authority is embedded in the legal system. This translates into hierarchical, coercive institutions similar to those of existing states, although legal communities outside states may be able to provide similar rule-of-law protections.

Is it possible that nonconsensual authority that is rule-of-law compatible develops in anarchist systems? I suppose it is a logical possibility. Pluralist, nonanarchist systems could be nonconsensual in the right way to ensure generality and remediability. It is just that in practice, and predictably, communities tend to ensure rule-of-law compliance via mechanisms that involve a monopoly of force. While nonconsensual authority that ensures compliance with the rule of law could be in principle polycentric, it is less likely to work because it requires effective and fluent coordination between polycentric agents that monopolize the use of force in their respective domains or over their legal subjects.

The anarchists will point out that existing states also fall well short of realizing the rule-of-law ideal, including equality before the law and accountability for violations of basic rules. People with more resources or access to legal expertise avail themselves more easily of the protections inherent in the law in countries that uphold formal rule-of-law values, and the barriers to certain groups such as women and minorities can be quite significant. And they would be right to think in terms of comparative institutional analysis. But it would be wrong for anarchists to take my argument as another instance of the nirvana fallacy, namely, of comparing law under anarchism with the perfect state-based legal system.⁵⁵ Although it is true that states fail to uphold rule-of-law values for a variety of reasons, the various shortcomings could in principle be remedied. The real shortcomings of legal systems within states can be addressed and have been addressed in liberal democracies, to create incrementally better rule-of-law protections, and as such they are not structural failings. By contrast, the structural challenges of consent-based legal systems show that rule-of-law ideals are not likely to be realized under anarchism, unless they are the product of deliberate design. It makes a difference that modern liberal democratic states place the rule-of-law ideal at the core of their legal systems while many anarchist systems place consent at their core.

While I have highlighted structural problems of anarchist legal systems, I have not made an all-things-considered case against anarchism. It is possible that states have other deficiencies that make them comparatively less desirable as forms of political organization. Furthermore, one can imagine situations in which anarchist legal systems must be respected as providing the only law available: international politics may be one such case. Failed states and societies marked by intractable, violent conflict, in which centralized bureaucracies either have collapsed or are no longer able to provide a minimum of law and order, may be another. However, it is important to understand whether legal systems under anarchism fall short of essential rule-of-law ideals, so that comparative institutional analysis aimed at reform is guided by an accurate appreciation of their strengths and weaknesses.