The duality of shared responsibility

ABSTRACT

This paper examines the duality of shared responsibility. The paper argues that shared responsibility is generally regarded as having the potential to address responsibility gaps in situations of concerted action. As such, shared responsibility may be important for global governance in relation to such diverse areas as peace-keeping, climate change, migration, and conservation of natural resources. The paper also argues that the sharing of responsibility can lead to a diffusion of responsibility that makes it more difficult to determine who is responsible for what. Thereby, it can undermine the effectiveness of global governance and moreover generate a new set of responsibility gaps. However, on the basis of international law, principles can be articulated that operationalize the abstract idea of shared responsibility and mitigate these negative consequences.

KEYWORDS:

• International law
• international responsibility
• international organizations

This paper examines the duality of shared responsibility. The paper argues that shared responsibility is generally regarded as having the potential to address responsibility gaps in situations of concerted action. As such, shared responsibility may be important for global governance in relation to such diverse areas as peace-keeping, climate change, migration, and conservation of natural resources. However, the paper also argues that the sharing of responsibility can lead to a diffusion of responsibility that makes it more difficult to determine who is responsible for what. Thereby, it can undermine the effectiveness of global governance and moreover generate a new set of responsibility gaps. The paper explores principles and processes of responsibility that can operationalize the abstract idea of shared responsibility and mitigate these negative consequences.

The paper addresses the potentials and limits of shared responsibility and possible solutions to problems of diffusion primarily from the perspective of international law. Sharing responsibility can be understood and evaluated from many different perspectives. International law is one of these perspectives, since sharing responsibility is in large part a matter of construing international obligations and allocating legal responsibility when actors breach obligations and cause injury. This legal dimension of shared responsibility is of critical importance for structuring international governance in areas such as climate change, conservation of natural resources, and multinational military operations. The paper thus will draw on the literature on responsibility in international law and on the codification of this body of law by the International Law Commission ([ILC], 2001, 2011). However, shared responsibility is a concept that can usefully be understood and studied from the perspective of other disciplines, notably political science and moral philosophy. Where relevant, the paper will refer to insights from bodies of literature other than international law.

I will first set out the dynamics that lead to situations of concerted action in relation to which the principle of shared responsibility is commonly applied (section 1). I will then introduce the dominant principle of individual responsibility and explain its shortcomings in situations of concerted action (section 2). In section 3 I will introduce the ambition and promises of the concept of shared responsibility as an antidote to the limitations of individual responsibility. I will then explore how shared responsibility may result in diffusion of responsibility and blame games, which may not only undermine its promise to address responsibility gaps but also generate new gaps (section 4). In section 5 I will discuss possible antidotes to the problem of diffusion of responsibility. Section 6 concludes.

1. Underlying dynamics

The promises and the downsides of shared responsibility arise both in cases of concerted action that may result in harm to third parties, and where actors independently cause harm (e.g. climate change) (Nollkaemper & Jacobs, 2013, pp. 370–372). In this paper I will focus on the former category.

To say that actors engage in a concerted action means that they jointly coordinate, plan, or carry out a particular action with a view to achieving a particular outcome. Concerted action involves some form of coordination between participating actors. This may be in the form of agreement, but mostly is of a more informal nature (May, 1996, p. 41). By engaging in concerted action actors can bring about results that they could not have brought about on their own (Erskine, 2008, 2015; May, 1996, p. 41).

Concerted action is central to global governance. It reflects the passage from a ‘society’ mainly characterized by coexistence to one also characterized by cooperation (Friedman, 1969, p. 127; Petersmann, 2011, p. 23). In areas where states have become dependent on each other to pursue common goods, they have felt compelled to address them jointly. In some cases, factual effects extend across borders and create interdependencies; examples are transboundary environmental effects, depletion of natural resources, trade in endangered species, piracy, refugee flows, human trafficking, the arms trade, and transboundary crime. In other cases changed perceptions create interdependencies; the recognition that it is no longer acceptable that genocide or mass killings are committed within a particular state is an example.

While cooperation and concerted action triggered by such interdependence is commonly intended to address harmful situations, it also may result in harmful outcomes. Sometimes such outcomes are intended; in other cases they are unintended side effects (Buchanan & Keohane, 2009; Harlow, 2011, p. 173). Examples can be drawn from many issue-areas. States that cooperate to conserve fish stocks beyond their Exclusive Economic Zone often have failed to realize agreed objectives, such as those in the 1995 ‘Agreement Relating to the Conservation and Management of Straddling Fish Stocks and Highly Migratory Fish Stocks’. In response to such failures, attempts have been made to clarify the responsibilities of multiple actors engaged in illegal fisheries (for example via the ‘Request for an Advisory Opinion Submitted by the Sub-Regional Fisheries Commission’ of 2013). Joint military operations have, not infrequently, resulted in the loss of civilian lives. Cross-border police operations may result in violations of individual rights. And states that engage in collective action to protect the climate may, individually and collectively, fail to achieve agreed targets.

The fact that states, international institutions and, increasingly, other actors cooperate in response to perceived common problems proportionately increases the situations where harmful outcomes may result from such concerted action. It is in these situations, where states and other actors come together to address common problems, and as a result of the concerted action contribute to harm, that both the potential and the downsides of shared responsibility emerge.

2. The principle of independent responsibility

The dominant approach to international responsibility has been the principle of ‘individual’ or ‘independent’ responsibility of states and international organizations. Under this principle, actors are only responsible for their own wrong, irrespective of their connection to other actors (ILC, 2001, Art. 47, para. 8).

The dominant paradigm of individual responsibility rests on two grounds. The first ground is of a methodological nature. For the purposes of determinations of responsibility, the relevant units of analysis would be individual actors rather than collectivities of actors. This ground is premised on methodological individualism: in the final analysis all conduct is explained by the actions of individuals (Arrow, 1994; Lukes, 1968, p. 119; Watkins, 1952, p. 186). This approach is commonly applied to justify individual criminal responsibility (Van der Wilt, 2005, pp. 91–108). Given the centrality of the sovereign state as the dominant agent in international law, this perspective can also be applied to justify the focus on individual states as agents that cause harmful effects.

The second ground underlying the paradigm of individual responsibility is of a normative nature. Just as international criminal law rejects the concepts of collective responsibility or guilt by association, instead relying on the principle of individual autonomy to limit responsibility to individuals only for their actual conduct, it would be normatively problematic to require states to be responsible for conduct other than their own (Case Concerning the Application of the Convention on the Prevention and Punishment of the Crime of Genocide [Genocide Case], ICJ, 2007, para. 406).

This idea of independent responsibility is reflected in the Articles on the Responsibility of States for Internationally Wrongful Acts (ARSIWA). The idea that individual states commit individual wrongful acts underlies the ARSIWA as a whole. The ILC was of the opinion that situations in which responsibility was attributed to a state other than the state that committed the internationally wrongful act were exceptional, and state responsibility for the state's own wrongful conduct came to be the basic rule underlying the ARSIWA.

International courts have adopted and reinforced this principle of independent responsibility. The International Court of Justice (ICJ) focused on independent wrongdoing in the Corfu Channel Case (United Kingdom v. Albania [Corfu Channel Case], ICJ, 1949) (despite wrongdoing by Albania and, allegedly, Yugoslavia) and Certain Phosphate Lands in Nauru (Nauru v. Australia) [Nauru], ICJ, 1992) (despite wrongdoing by Australia and, allegedly, the United Kingdom and New Zealand). Likewise, the European Court of Human Rights (ECtHR) considered the responsibility of Belgium and Greece independently in M.S.S. v. Belgium and Greece (ECtHR, 2011). The Tribunal in the Eurotunnel Arbitration (The Channel Tunnel Group Ltd & France-Manche S.A. v. United Kingdom & France [Eurotunnel Arbitration], Permanent Court of Arbitration, 2007, para. 187) also approached international responsibility for the common conduct of France and the United Kingdom through the lens of independent responsibility.

It can be inferred from these cases, in particular the Eurotunnel Arbitration and the East Timor Case (East Timor (Portugal v. Australia) [East Timor Case], ICJ, 1995), that individual actors retain their individual obligations, even when they act in concert. Moreover, in principle, the fact that more than one actor is engaged in a particular wrongful act does not release each individual actor from its responsibilities. In the East Timor Case Judge Weeramantry, dissenting from the majority judgment, noted that ‘[e]ven if the responsibility of Indonesia is the prime source, from which Australia's responsibility derives as a consequence, Australia cannot divert responsibility from itself by pointing to that primary responsibility’ (East Timor Case, ICJ, 1995, Diss. para. 172). Australia's own role in regard to the treaty was therefore sufficient for its responsibility – irrespective of the responsibility of any other actor.

However, while the principle of independent responsibility has thus been applied to situations of concerted action, it has significant shortcomings. In combination with the procedural limitations of dispute settlement (international courts may find that they do not have jurisdiction to hear a case unless all responsible parties are before the court), the concept of independent responsibility has led courts to reduce complex cooperative schemes to binary categories without engaging in principled discussions of the shared nature of responsibility (Nauru, ICJ, 1992; Corfu Channel Case, ICJ, 1949; East Timor Case, ICJ, 1995; Legality of the Use of Force (Serbia and Montenegro v. United Kingdom) [Legality of the Use of Force], ICJ, 1999; Military and Paramilitary Activities in and Against Nicaragua (Nicaragua v. United States) [Nicaragua], ICJ, 1986). A noteworthy example is the decision of the ECtHR in Behrami (Behrami v. France [Behrami], ECtHR, 2007). The Court attributed all acts and omissions relating to the failed demining operations in Kosovo exclusively to the United Nations, and not to its member states, without considering the possibility of a more nuanced solution in which responsibility would be shared.

A system of responsibility that disaggregates responsibility into individual cases of wrongdoing, and that does not connect well to the structure of international cooperation, may undermine key benefits of the law of responsibility. It may hinder answerability for the exercise of public authority, sustain collective action problems, and legitimize harmful practices. Its most visible drawback is that it makes it difficult for persons who suffer injury to figure out who is to blame for the harmful effects. Attributing responsibility only to one actor, even though another actor contributed to the harmful outcome, also raises further questions. For instance, if only one state is held responsible and that state cannot compensate all injury, is it fair to leave the injured party with the remaining costs? Is it fair to hold the single state responsible for all injury? These questions were raised, but not answered, in Nauru, ICJ, 1992. Australia, New Zealand, and the United Kingdom were designated as the joint authority to exercise the administration of Nauru. As a result of their administration, Nauru's natural resources were exhausted. If Australia alone had been held responsible, the question would have been whether it would be fair to require Australia to pay all the damages to Nauru, and leave the United Kingdom and New Zealand ‘off the hook’.

In sum, independent responsibility does not appear to be conducive to a rule-based society in which responsibility fulfills the essential function of ensuring a return to legality and ensuring that the actors that acted in breach of international law will comply with their obligations (Brownlie, 1998, p. 79). Reducing situations of shared responsibility to individual responsibility may create an accountability gap that implies costs for the injured parties and the larger system, and raises questions of fairness among the responsible parties.

3. The promise of shared responsibility

The principle of shared responsibility seeks to mend the shortcomings of individual responsibility. I use the concept of shared responsibility to refer to situations where a multiplicity of actors contributes to a single harmful outcome, and legal responsibility for this harmful outcome is distributed among more than one of the contributing actors (see for the use of the term ‘harmful outcome’ e.g. Cane, 2001, p. 88; Honoré, 1999, p. 27; Miller, 2007, p. 383; Stern, 2005, p. 93).

The defining feature of shared responsibility is thus that the responsibility is allocated to two or more actors that work together with a view to achieving a particular outcome. This aspect of shared responsibility builds on the notion of collective responsibility developed by Erskine (2008, 2015). However, responsibility is distributed between actors separately, rather than rested on them collectively (May, 1996, p. 112). If responsibility were to rest on a collectivity, it would no longer be shared, but would instead be a responsibility of the collectivity as such (May, 1996, p. 116). Somewhat counterintuitively, shared responsibility by definition thus rests on individual actors for their contribution to a particular harm.

However, shared responsibility is not simply the aggregation of two or more individual responsibilities. In many situations of shared responsibility, two or more actors stand in some relationship to each other, and their conduct or omissions mutually influence the scope of responsibility of the other. An important application of the concept of shared responsibility is to situations where responsibility is based on multiple actors contributing to each other's acts and thereby to the eventual outcome.

This ‘relational account’ of international responsibility is based, on the one hand, on interdependence of conduct and, on the other, interdependence of outcomes. As to the former, the conduct of one state or international institution is conditional on and/or conducive to acts or omissions of other actors that are part of the concerted action (Lucas, 1995, p. 75). If cooperative conduct cannot be reduced to the conduct of individual participating actors, responsibility needs to connect to the relationship between the individual actors. Individualizing responsibility may miss the point. As to the latter, concerted action can achieve results that could not be realized by actors acting alone (May, 1987, p. 26). Erskine (2015, p. 256) notes that ‘agents who come together, even in an informal association, to work towards a shared goal are able to achieve things by cooperating that they would not be able to achieve independently’. Extraordinary rendition is a good example. The conduct of states like Macedonia and Poland was influenced by the conduct of the United States, and in turn influenced the subsequent conduct of the United States (El-Masri v. the Former Yugoslav Republic of Macedonia [El-Masri], ECtHR, 2012, p. 263). Together they realized results that they could not have achieved alone. Erskine noted that the very ability of actors to achieve results that they cannot achieve alone can even be the basis for a (moral) obligation to work together (or, in my terms, to engage in concerted action) (Erskine, 2015). When states act together, they rely on each other and can achieve together more than they can alone; this is a basis for the distribution of responsibility between all of them.

The principle of shared responsibility thus defined can provide important benefits to international governance. A system where international law distributes legal responsibility to multiple actors who contribute to a single harmful outcome provides better incentives for rule-conforming behavior and can better protect the interests of victims in situations where states and other actors act in concert rather than alone. In a world where each actor acts on his own, and harm arises out of individual conduct, individual responsibility could suffice. In contrast, when actors frequently act together and cause harm together, a concept of responsibility that reflects that ‘togetherness’ is appropriate.

4. The downside of shared responsibility

Notwithstanding the potential benefits of shared responsibility to close responsibility gaps, the general notion of shared responsibility can also result in new gaps. The consequence of the fact that responsibility is spread over a multitude of persons may be that the actual share of responsibility of each person involved becomes smaller (Linklater, 2011, p. 225). This may result in a ‘diffusion of responsibility’, well captured in Bovens’ observation that ‘[a]s the responsibility for any given instance of conduct is scattered among more people, the discrete responsibility of every individual diminishes proportionately’ (Bovens, 1998, p. 46). Diffusion of responsibility in cases of concerted action is a manifestation of the so-called ‘problem of many hands’ (Thompson, 1980). Thompson argued that assigning responsibility in the framework of governmental organizations becomes more difficult when more persons – ‘many hands’ – are involved in the process that caused harm. The problem of many hands is relevant in an international context.

The problem of diffusion of responsibility can be further explored by distinguishing between causes of diffusion, designed diffusion, and costs of diffusion.

4.1. Causes of diffusion

Diffusion of responsibility is essentially a normative phenomenon. The problem of many hands arises from the fact that it is morally problematic to attribute responsibility to individuals where that could not be justified on moral grounds (Thompson, 2014; Van de Poel, Fahlquist, Doorn, Zwart, & Royakkers, 2012, p. 64). This applies equally in the legal domain. Diffusion of responsibility arises when responsibility cannot be assigned or determined because the legal conditions for responsibility have not been (fully) met (ILC, 2001, Arts. 1–2, 2011, Arts. 1–2).

The difficulty of identifying who is responsible for a harmful outcome in a collective setting may then be due to the fact that individual contributions may be too small or otherwise insufficient to meet criteria for responsibility. In situations of many hands, tasks may be divided up, so that multiple actors perform small tasks which combine to result in larger (harmful) outcomes (May, 1996, pp. 7–8, 73). Individual actors then might not meet the conditions for responsibility. Whether or not this is a situation that will frequently occur is perhaps uncertain. No systematic empirical work has been carried out. However, a review of multiple areas of practice in such diverse areas as military operations, fisheries, transboundary watercourses, and pollution of the marine environment does suggest that harmful outcomes originating from a cumulation or interaction of multiple smaller causes is in fact quite common (Nollkaemper & Plakokefalos, 2016).

Two specific situations demonstrate where it is difficult to identify who is responsible for what. The first situation arises when obligations are assigned to a collectivity, rather than to particular actors (‘unassigned obligations’). This is particularly relevant for obligations to prevent (Naftali, 2009). Such obligations may allow states to escape their responsibility by pointing to the non-performance of obligations by others (Dupuy, 1999, p. 371; Economides, 2010, p. 371; Van de Poel et al., 2012, p. 52). David Miller's observation that ‘an undistributed duty  …  to which everybody is subject is likely to be discharged by nobody unless it can be allocated in some way’ is relevant for diffused responsibility (Miller, 2007, pp. 99–100). For example, the principle of ‘responsibility to protect’ assigns obligations to protect human populations from mass atrocities to a plurality of states or organizations (Amnéus, 2008, pp. 502–25; Hakimi, 2010, pp. 342–43; Vetlesen, 2000, p. 529) – it may then be difficult to determine which of the actors indeed had to act. While this example does not directly concern responsibility in the way it is discussed in this paper, but rather reflects that the obligations of states may leave who is to do what unclear in the first place, this has direct implications for responsibility. This was so in the Genocide Case before the ICJ: because it was unclear which states had to do what, the Court was not able to conclude that Serbia had to provide compensation to Bosnia and Herzegovina for its own failure to act.

A second cause of diffusion of responsibility is that ‘many hands’ make it difficult to identify who did what. For outsiders, ‘it is usually very difficult, if not impossible, to know who contributed to, or could have prevented, a certain action, who knew or could have known what’ (Van de Poel et al., 2012, p. 61). It may not be clear ‘even to the members of the collective itself who is accountable’ (Nissenbaum, 1996, p. 29). So conceived, the problem of many hands is an epistemological problem: the problem of identifying who is responsible for what arises from a lack of knowledge or information (Van de Poel et al., 2012, p. 61). This practical problem arose in the case brought by Serbia against multiple states of the North Atlantic Treaty Organization (NATO) (Legality of the Use of Force, ICJ, 2004, para. 279) and in the Saddam Hussein Case (ECtHR, 2006, para. 1), in which Saddam Hussein brought a case against twenty-one states that were allegedly implicated in the invasion of Iraq and where it was next to impossible for the plaintiffs to identify who did what. It is also well illustrated by the fact that fifty-four states participated in the United States’ rendition policy (Council of Europe [CoE], 2007; Hammarberg, 2011; Open Society Justice Initiative [OSJI], 2013).

This problem is exacerbated by the informal nature of arrangements in collaborative action. Examples are joint action on piracy (where rules of engagement will usually be beyond the reach of plaintiffs) and cross-border joint policy operations; transborder police cooperation (Hufnagel, 2013); financial arrangements within the Basel Committee on Banking Supervision (Zaring, 2004, p. 547); the rules of the Nuclear Suppliers Group; and command and control structures in military operations (Tondini, 2010).

4.2. Designing diffusion

Diffusion is not a phenomenon that is ‘out there’, but rather results from the conduct and strategies of relevant actors. Diffusion may be ‘orchestrated’ (Abbott, 2014). Orchestration need not (and commonly will not) be expressly aimed to achieve diffusion, but this may be the result. In particular cases, however, the result may well be intended. Diffusion may be designed at three different levels: by setting up concerted action as such; by the way tasks are allocated among participating actors; and by principles of responsibility.

As to the first level, the very choice for concerted action may diffuse responsibility. From the perspective of the participants in a particular concerted action, adding more ‘hands’ can be necessary for the success of the cooperation (for instance in the case of climate change, where significant reductions of greenhouse gases cannot be achieved if the number of cooperating states is low). However, adding more partners also may increase the possibility of diffusion, and thus may shield participants from responsibility. Partnerships between international institutions and other actors are a good example. Partnerships engaged by the World Health Organization (WHO) or the World Bank involve multiple actors. When no clear arrangements have been made on the assignment of responsibility, it becomes difficult to determine who is responsible for what (Boisson de Chazournes, 2013, pp. 215–16; Clarke, 2011, p. 55, 65).

At the second level, diffusion may result from the allocation of tasks within a concerted action. Actors may exploit the notion of sharing as a way to diffuse responsibility. It may be the intention of actors that seek to diffuse responsibility to ensure that they themselves are protected from claims. Diffusion may result from blame games, ‘where multiple players are trying to pin the responsibility on one another for some adverse event, acting as blamers to avoid being blamees’ (Hood, 2011, p. 7).

Two strategies can be distinguished. One is to ‘blunt’ responsibility by collectivizing it (Hood, 2011, p. 100). Relevant actors spread powers and tasks over multiple actors, as a result of which the outside actors cannot easily identify who is to blame for any particular event. Partnerships between international institutions and other actors have precisely this effect. It may be unclear whether diffusion of responsibility and deflection of blame are the intended aims of such partnerships, or simply unintended consequences. However if no clear arrangements have been made on the assignment of responsibility, responsibility rests on all, and as a result, perhaps on no one. A comparison can be drawn with Jeremy Bentham's proposition that in view of the blame-spreading potential of collective decision-making, collective bodies could be seen as ‘screens’ (cited in Hood, 2011, p. 97). Such strategies make it unclear who did what ‘and leave potential blamers nonplussed by the complexity of the organizational arrangements’ (Hood, 2011, p. 83).

An alternative strategy is to individualize blame (Hood, 2011, p. 100). Rather than collectivizing blame, blame is then shifted to one or a few actors, in effect shielding the blamers. As noted by Hood (2011, p. 100), this strategy ‘is about shifting rather than reducing or preventing blame’ (see also Hood, 2002). In various proceedings relating to the genocide in Srebrenica in 1995, it was sought to hold the Netherlands and the United Nations (UN) responsible in relation to the eviction of persons from the UN compound in Srebrenica, both defendants denied responsibility and effectively placed the blame on each other (Nollkaemper, 2008). Such strategies can thus result in responsibility being assigned to one actor only. Delegation can have the same effect: by delegation, states and international institutions can act with and through others, with potential limiting effects on the scope of their own responsibility (Sarooshi, 2004, p. 291). Furthermore, states can ‘outsource’ tasks to other actors, such as private security corporations (Chesterman & Fisher, 2009; Seiberth, 2013). The outsourcing state then can act through others and accomplish objectives without running the risk of being held responsible.

At the third level, diffusion can be designed by way of principles of responsibility. The form and content of any scheme of responsibility are a matter of conscious design, mostly by states or international organizations but also by international courts. Both rules on responsibility and primary rules of conduct pertaining to, say, environmental law, trade law, or military matters are a result of a political choice. The law of responsibility, and the institutions and processes in which it is embedded, is in itself the result of choices and practices of states. The normative and institutional choices reflect the ‘productive power’ of the relevant actors over others (Barnett & Duvall, 2005, p. 18). States exercise power over international law-making or particular international institutions, and by influencing the rules of responsibility they can determine whether or not the exercise of a particular type of conduct (including participation in concerted action) engages the responsibility of a state. The ineptitude of international law for dealing with harmful consequences of concerted action serves states and other actors well, by allowing them to engage in blame-avoidance and blame-shifting, thus shielding themselves from responsibility. Indeed, the fact that the law of responsibility has been designed in such a way that it results in exclusive, rather than shared, responsibility (as noted in section 2) is precisely the result of such political choices.

4.3. Costs of diffusion

While diffusion in itself is a neutral term that frames and describes the spread of ideas, institutions or, as in the case of responsibility, legal principles and processes, diffusion in the sense discussed above can significantly undermine the potential of shared responsibility to close the gaps created by the principle of individual responsibility. This is so in three respects: in terms of accountability, in terms of the performance of obligations, and in terms of injured parties.

First, diffusion undermines accountability. Responsibility as a legal concept is one subset of the wider, non-legal concept of accountability (Bovens, 2007, p. 447; Brunnée, 2005; Curtin & Nollkaemper, 2005). In this sense responsibility is part of our expectations of how public authority is exercised. It belongs in the same basket of terms such as legitimacy, transparency, democracy and, more generally, good governance (Bovens, 2007, p. 447). Perhaps the most relevant dimension of responsibility in this sense is answerability (Crawford & Watkins, 2010, p. 283; Gardner, 2003, p. 157; Lucas, 1995). Diffusion can undermine this feature of accountability. That is, it may be unclear who has to answer for what. Illustrative examples are the accountability problems in the operation of multinational military operations (Zwanenburg, 2005) and those arising from the global financial crisis (Miller, 2015).

Second, diffusion of responsibility can undermine the assignment and performance of obligations and thereby the achievement of objectives. It is a plausible proposition that diffusion of responsibility can undermine incentives for action. If no one can be meaningfully held responsible after the event, this may impact on the willingness of an actor to feel obliged to act beforehand (Bovens, 1998, p. 49; Wallach, Kogan, & Bem, 1964, p. 263). This may reduce the possibility that individual actors perform their obligations and that the interests that the law seeks to protect are actually protected.

Whether diffusion will indeed undermine the incentives of actors to perform their obligations presumes, as a first step, that obligations matter at all for the conduct of the relevant actors (Posner & Sykes, 2013) and, as an extension, that the perspective of being held responsible is a relevant factor in changing the conduct of states and international institutions. This connection is not implausible. Responsibility may strengthen the internalization of obligations – it is one of the main factors that support compliance with international obligations (Jinks, 2004, p. 622). Responsibility may also impact on the reputational impact of international law (Guzman, 2006, p. 379). States may care about the reputational effects of non-compliance, but might do so even more when such non-compliance may trigger their international responsibility.

This impact of diffusion on incentives may have wider ramification in the form of collective action problems. This will be particularly relevant when the participation of multiple states is necessary to address a perceived problem and to produce a common good, for instance in situations involving transborder effects in areas such as global health, financial markets, the environment, or organized crime, where any single state is quite powerless to provide answers (Allum & Gilmour, 2012, pp. 1–2; Mason, 2008, p. 8; Mishkin, 2011, p. 68). Precisely because obligations and responsibilities are not specifically assigned, and responsibility is not likely to be forthcoming, actors may be inclined to wait for each other to act, with the result that nothing happens. Olson, who developed the theory in the economic context of public goods, argued that as members of a large group generally hold the assumption that someone else in the group can and will provide the public good, the incentives for these members to provide it themselves are weakened (Olson, 1968, p. 44). Diffusion of responsibility may strengthen this phenomenon. If persons do not individually experience the consequences in terms of being held responsible, members may be tempted to look for others to do the job, or simply not care (Hardin, 1971, p. 472, 479).

Third, diffusion of responsibility adversely affects the position of injured parties. In addition to the role of responsibility in creating incentives for action, one of the central aims of responsibility is to provide redress to injured parties. As such, it provides a key criterion for evaluating the impact of diffusion of responsibility. Prima facie, it would seem that diffusion can undermine a key function of attributing responsibility: to ensure justice to victims (Bovens, 1998, p. 49; Van de Poel et al., 2012). If harm is caused, responsibility cannot be determined, and injured parties will be without redress (Stöckle, 2013). In effect, the loss will then be left where it falls – with the victim, rather than being transferred back to one or more responsible actors (Crawford & Watkins, 2010, p. 286).

One reason why the position of victims tends to be weaker in situations of concerted action is that it may be more difficult for private parties to determine which actors played what role in a particular concerted action. If contributions are spread among multiple actors, the relative contributions of individual actors will be relatively small. Moreover, information is disseminated among many actors and is of an informal nature – for instance in the case of partnerships between international institutions and private parties (Boisson de Chazournes, 2013, p. 211).

It should be added that it would overly simplistic to see the problem of many hands only as a problem that should be removed. The permissiveness of the law (that allows conduct that contributes to harm without triggering correlating responsibility) may also induce states to agree and engage in cooperative action intentionally. For some, or all, of the actors that are involved in a coordinated effort, the difficulty of determining responsibility may actually be an advantage and might well be a condition for engaging in cooperation to protect in the first place. In terms of a cost–benefit assessment, dilution of responsibility may then even be beneficial: action without responsibility may produce better outcomes than no action at all.

Nonetheless, the scenario where cooperative action may result in diffusion of responsibility that sits uneasily with the promise of shared responsibility is a real one. It is with regard to that scenario that the next section considers how the principle of shared responsibility may be operationalized and made effective after all.

5. Antidotes

It follows from the above that the idea of shared responsibility, while potentially conducive to effective global governance, can lead to problems of diffusion of responsibility, in particular because it may be unclear who has to do what and, ex post facto, who of a multitude of actors was responsible for a particular harm. If shared responsibility is to live up to its promise and ambition, it will be necessary to operationalize it in more specific principles and processes.

In part, such operationalization will take the form of negotiation of specific rights or obligations, or setting up of international institutions that address the underlying problems. The UN Framework Convention on Climate Change and the ensuing negotiations, the Straddling Fish Stocks Agreement, and the Arms Trade Treaty are examples of regimes that specify in advance who has to do what, and thereby also who will be responsible for what in case of injury, and thus may prevent problems of diffusion. They are examples of ‘mediating institutions’ that stand between, and connect, duty bearers and rights holders (Shue, 1988).

For another part, the problems of many hands may be counteracted by monitoring and supervision arrangements that make it possible to identify the contribution of each actor. The detailed reports compiled within the framework of the Convention on International Trade in Endangered Species (CITES) on the roles and infractions by individual parties who, collectively, contribute to the extinction of particular species, is an example. However there are considerable differences in the existence of such mechanisms and their ability to obtain the relevant information, with direct consequences for diffusion of responsibility.

I will leave both the ‘ex ante’ responses to potential problems of many hands and the use of monitoring mechanisms aside, and instead focus, in line with the approach of the paper, on principles of responsibility that may alleviate the problems of many hands in situations where such regimes either have not been negotiated or, alternatively, do not achieve the intended consequences.

I will focus on three principles of responsibility that can be viewed as operationalization of shared responsibility and that may serve as antidotes to the problems of diffusion of responsibility. The common feature of these principles is that each would provide a basis for responsibility of all actors who have contributed to a particular harm. Thereby, they directly counteract the fundamental cause of diffusion that was identified above: the fact that individual contributions may be too small to meet criteria for responsibility, even though they result in larger (harmful) outcomes (May, 1996, pp. 7–8, 73). For each of these principles I will assess in what ways the principle may alleviate the problems of diffusion, as well as the basis of these principles in international law.

It should be noted at the outset that all three principles fall short of the principle of responsibility that intuitively might be thought to be the most effective response to the problems of diffusion: the principle of joint and several responsibility. According to this principle, each actor who has contributed to harm is responsible for the acts of the others (‘joint’ responsibility) and may be individually asked to make full reparation (‘several’ responsibility) (Crawford, 2000, p. 74, para. 272). This is the archetypical solution to situations where two or more actors contribute to damage and it is unclear what part of the damage is caused by whom (Alford, 2011; Noyes & Smith, 1988). Proposals for the introduction of this principle in international law are inspired by domestic law, where joint (or ‘joint and several’) liability is frequently used to solve liability questions involving multiple tortfeasors (Principles of European Tort Law, Art. 9:101(2)). In international law, the principle is contained in some treaties (the Outer Space Liability Convention, Art. IV; the United Nations Convention on the Law of the Sea, Art. 139). The Seabed Chamber affirmed the applicability of this principle under the Law of the Sea Convention, writing ‘[j]oint and several liability arises where different entities have contributed to the same damage so that full reparation can be claimed from all or any of them’ (Responsibilities and Obligations of States Sponsoring Persons and Entities with Respect to Activities in the Area, 2011, para. 201). However, apart from treaty provisions, the principle is not part of general international law (D’Argent, 2014; Wittich, 2009). More importantly, it is a compelling argument that the necessary conditions for its implementation are absent in international law. In particular, the principle of joint and several responsibility presumes the presence of a compulsory court system that can adjudicate claims when one wrongdoing actor seeks recourse from another responsible actor (D’Argent, 2014).

Rather than transposing the principle of joint and several liability to international law, this section identifies three more limited principles that may have a more realistic prospect for implementation in international law: the principle of shared responsibility for the same wrongful act; the principle of shared responsibility for cumulative harm; and the principle of shared responsibility for concerted action.

5.1. Shared responsibility for the same wrongful act

Of these three antidotes, the most firmly established is the principle that two or more states and/or international organizations that commit the same wrongful act are both responsible for the consequences. The notion of the ‘same wrongful act’ implies that two or more actors engage in ‘a single course of conduct’ (ARSIWA Commentary, p. 124, para. 3). This principle would apply in the situation where a common organ of several states carries out a wrongful act (Crawford, 2002, p. 272). An example is the Coalition Provisional Authority set up by the United Kingdom and the United States during the occupation of Iraq (Talmon, 2008). Regarding this example, the ILC noted that

the conduct of the common organ cannot be considered otherwise than as an act of each of the States whose common organ it is. If that conduct is not in conformity with an international obligation, then two or more States will concurrently have committed separate, although identical, internationally wrongful acts. (ILC Yearbook, 1978, p. 99)

In such a situation, the simple principle applies that each actor is separately responsible for the conduct attributed to it (Crawford, 2002, p. 272). Since the conduct can be attributed to multiple actors, responsibility necessarily is shared: ‘the responsibility of each State [or organization] may be invoked in relation to that [same internationally wrongful act]’ (ARSIWA, Art. 47).

This principle is an example of the possibility that a single course of conduct can be attributed twice (so-called dual attribution: Condorelli, 1995, p. 881, 1997, p. 87), a possibility that has been recognized particularly in the context of peace-keeping operations (Dannenbaum, 2010, p. 113; Sari, 2008, p. 151) in relation to, for example, the possibility of multiple attribution of conduct to both international organizations and troop-contributing states (HN v. Netherlands (Ministry of Defence and Ministry of Foreign Affairs) [HN v. Netherlands], DCTH, 2008, paras. 47–49; Nollkaemper, 2011, p. 1143; Nuhanović v. Netherlands, THCA, 2011; Tsagourias, 2011).

The principle of shared responsibility for acts of common organs can also be supported in another way. It follows from the fact that if one act can be attributed to multiple persons, there is only one cause of the harm. That cause then cannot be apportioned or divided between multiple actors (D’Argent, 2014). Because all the actors committed the act, and the contribution of none of them can be separated, each actor will be responsible for the whole outcome.

The outcome whereby each actor is responsible for the whole injury looks similar to the principle of joint and several responsibility, to which reference was made above. However, it cannot be taken as a basis for the conclusion that a wider principle of joint and several responsibility applies in international law. In fact, it is rare for states to act through common organs, and there are only a limited number of examples where questions of shared responsibility in such contexts have been raised. The principle thus certainly cannot be transposed to other situations where states do not act through a common organ. Moreover, the principle itself is not accompanied by principles and procedures that are typically part of joint and several liability schemes: in particular, the principle that a responsible state can claim any paid compensation from another responsible actor.

5.2. Shared responsibility based on cumulation

The second antidote to diffusion of responsibility is the principle that multiple actors share responsibility when all of them commit a separate (which distinguishes the principle from the principle discussed above) wrongful act, and the separate wrongful acts combine, or cumulate, in a single harmful outcome. One example is the Corfu Channel Case, where Yugoslavia committed a wrongful act of laying mines, which then combined with Albania's wrongful act of not notifying a third state, resulting in harm to the United Kingdom (Corfu Channel Case, 1949). D’Argent provides the scenario of a state, where an abduction took place, being instrumental in aiding or assisting that abduction by agents of another state. The wrongful act of aiding or assisting the abduction is a ‘cumulative’ cause of the illegal abduction committed by another state, since ‘the aid or assistance does not produce any harmful outcome severable from the abduction itself, but was necessary for it to happen’ (D’Argent, 2014, p. 228). None of the separate wrongful acts are by themselves sufficient to produce the harmful outcome – but in combination, they do cause that harmful outcome (D’Argent, 2014, p. 227).

Similar to the previous situation, in this scenario it will not be possible to identify separate causal connections between the harm and individual contributions – the various contributions form one undivided whole. In principle, therefore, this would be a situation where the risks of diffusion would arise. That risk is only counteracted by a principle whereby all states contributing to the harm are responsible.

Whether or not this principle exists as a matter of positive international law is somewhat doubtful. Some argue that in these cases it would after all be possible to identify a single cause as the most important cause. This probably would be Albania in the example of the Corfu Channel Case. It would then only be that state that would be responsible. This solution would thus transform a problem of many hands in fact to a situation of a single hand, and thereby prevent diffusion of responsibility. This implies that this would not qualify as a situation of shared responsibility.

The alternative approach, which would provide for shared responsibility, would consider all of the contributing causes as equivalent causes – as causes without which the injury would not have occurred (D’Argent, 2014, p. 230). Since no single contribution could be singled out, all contributing actors would be responsible. This approach would preclude the drawbacks of diffusion as identified in section 4. It would offer the injured parties a choice to hold each participating actor responsible. Moreover, each of the contributing actors would be responsible, and thus would be required to return to lawful behavior.

Some support for this second interpretation can be inferred from the judgment of the ICJ in the Corfu Channel Case. While it was alleged that Yugoslavia had laid the mines (which certainly would be one cause of the harm), the Court found Albania responsible for the explosions and for the damage and loss of human life that resulted from them (Corfu Channel Case, 1949). It may be argued that both states were responsible for the full damage (Crawford, 2000, p. 19, para. 34) – this would indeed be consistent with shared responsibility.

It should be emphasized that while the principle thus has some basis in international law, it would only apply in rather exceptional cases. It only applies when multiple states in fact commit separate wrongs and these wrongs cumulate in a single harm. It would not apply in situations of armed attacks by coalitions, where multiple states collaborate and one of them commits a wrong, or in other situations of concerted action where not all of the actors commit a wrong, yet all of them contribute to the injury. For these situations we have to consider a third and less well-established principle.

5.3. Responsibility based on participation in concerted action

The third principle that would form an antidote to diffusion is the principle that the mere fact of participating in a concerted action that results in injury would be a sufficient basis for responsibility. The justification for this wider approach is precisely based on the factors that, as discussed in section 3, justify shared responsibility in the first place: shared responsibility reflects a ‘relational account’ of international responsibility that is based, on the one hand, on the fact that the conduct of one state or international institution is conditional on and/or conducive to acts or omissions of other actors, and on the other hand, on the fact that concerted action can achieve results that could not be achieved by actors acting alone (Erskine, 2015; May, 1987, p. 26). It is not only the contribution to injury as such that is the normative basis for shared responsibility, but rather the interdependence of action and outcomes of the concerted action. Moreover, this wider choice would be a political choice, based on protection of the interests of the injured parties and perhaps the hope to create incentives for rule-abiding behavior.

The principle of shared responsibility for participation in concerted action deviates in one fundamental sense from the two principles discussed above. For each of the above cases, all states or organizations involved would be held responsible for their own conduct. It is the conduct (leading to the harm) that is attributed to multiple states, which on that basis can be held responsible. In the cases of responsibility for concerted action the situation is different: in this case the responsibility is attributed to each of the participating states – even for conduct that was not one's own. The distinction between attribution of conduct and attribution of responsibility is well-established – though not uncontroversial – in international law (Fry, 2014). As such, responsibility based on concerted action is part of a wider category that also includes responsibility based on direction and control, delegation, or outsourcing.

Whether this principle is part of positive international law is more uncertain than the principles identified above. One can find inspiration in the use of concepts of joint enterprise as developed in international criminal law (Nollkaemper & Jacobs, 2013). There is some support for this principle in the ARSIWA Commentary, which refers to the situation where ‘two or more States (…) combine in carrying out together an internationally wrongful act in circumstances where they may be regarded as acting jointly in respect of the entire operation’ (ARSIWA Commentary, p. 124, para. 2). However, there are few, if any, examples where the principle has been accepted by international courts.

The principle could have come up in two cases before international courts. In the Legality of the Use of Force cases before the ICJ, Serbia alleged that multiple NATO states, and NATO itself, had a hand in the armed attack on the former Yugoslavia (Legality of Use of Force (Yugoslavia v. United States) p. 916 (9)). It could be said that given the role of multiple states and NATO itself, responsibility could be spread over all of them, rather than one state alone. However, the case was never decided on the merits by the Court. The second is the Saddam Hussein case before the ECtHR. In this case, the plaintiffs brought claims against a multitude of (allegedly) responsible states that took part in the coalition against Iraq. The Court rejected the claims, in part because the plaintiffs could not demonstrate who did what. Arguably, recognition of a principle of shared responsibility based on concerted action would have precluded the need to identify causal relations between particular actors and particular consequences.

In the current state of international law, these examples are more demonstrative of the role that the principle could fulfill in preventing diffusion of responsibility and giving specific meaning to shared responsibility, rather than providing a firm basis for the principle. The facts that this principle is based on attribution of responsibility rather than conduct and that states have always preferred a responsibility system in which they are only responsible for their own conduct and not that of others (Genocide Case, ICJ, 2007) establish barriers against a firmer basis for this principle.

6. Conclusion

This paper has exposed how shared responsibility, often heralded as a solution to situations where multiple actors contribute to harmful outcomes, can create new responsibility problems by resulting in a diffusion of responsibility. Such diffusion, and the resulting responsibility gaps, is in part explained by international law itself, including the dominant paradigms of responsibility, which to some extent is designed to allow for diffusion and blame games. This process of diffusion should be understood as a political process, which is sustained and driven by international regulation, which in itself is a reflection of that political process.

One might think that the obvious answer to diffusion and responsibility gaps that emerge in situations of shared responsibility would lie in better regulation. However, regulation in a cooperative context can itself trigger new situations of diffusion. This suggests that we can speak of a cycle of diffusion. Acceptance of shared responsibility often leads to a regulatory cycle which can sustain, rather than solve, a diffusion of responsibility.

Diffusion may result in costs, in terms of the accountability of the exercise of public authority, in terms of performance of international obligations, and especially in terms of protection of the rights of injured parties. It is quite obvious that if diffusion is a necessary part, and indeed a precondition, of regulation, the benefits of regulation are to be considered too – even when such regulation results in diffusion of responsibility. How the benefits and costs will be evaluated depends on a contextual analysis and cannot be answered in the abstract. Nonetheless, there is little doubt that the scenario in which diffusion undermines either a return to legality and/or compensation for victims is a real one.

It is in respect of that scenario that the three antidotes described are a critical part of the operationalization and application of shared responsibility. Of these three principles, the principle of shared responsibility for the same wrongful act is the most firmly established, yet it is also the most narrow, as it basically applies only in cases of common organs. The principle based on cumulation is slightly broader, but less firmly established. And unsurprisingly, the principle based on concerted action, which has the greatest potential to counter risks of diffusion, is the least established in international law. The degree to which this principle will acquire a firmer basis in practice and law may well be the ultimate test of the acceptance, and viability, of shared responsibility in international law.

Disclosure statement

No potential conflict of interest was reported by the author.

ORCID

Andre Nollkaemper http://orcid.org/0000-0002-4586-4940

Notes on contributor

Andre Nollkaemper is Dean and Professor of Public International Law at the Faculty of Law of the University of Amsterdam. He is also external Legal Advisor to the Minister of Foreign Affairs of the Netherlands, Member of the Permanent Court of Arbitration, Member of the Institut de Droit International and Member of the Royal Netherlands Academy of Arts and Sciences.

Additional information

Funding

This work was supported by FP7 Ideas: European Research Council [grant number 249499].