The central topic addressed in this essay is the relationship of duties of justice towards refugees and the distribution of responsibility for the protection of refugees (with particular focus on the questions of whether doing one’s fair share of refugee protection entails that one has done all that is required to discharge one’s duty of justice towards refugees). I address this topic in relation to the institutional context of the current international refugee regime. The first part of the essay lays out the normatively salient features of the current refugee regime and the relationship between duties of justice to refugees and the fair distribution of refugee protection that it institutes. In the course of this discussion, I draw particular attention to the fact that it is an important feature of the current refugee regime that it does not limit the duty to protect refugees to doing one’s fair share of refugee protection. The second part of the essay considers whether this feature of the refugee regime is justifiable by addressing arguments concerning whether duties of justice are limited to doing one’s fair share of refugee protection. In particular, this section addresses the argument proposed by David Miller that there is no duty of justice to take up the slack that is produced by the non-compliance of other actors. It argues that Miller’s argument is invalid and that duties of justice towards refugees may require doing more than one’s fair share, but also that there is an obligation to seek the fairest arrangements compatible with effective refugee protection. In the final section, I consider the significance and difficulties of the issue of a fair distribution of refugee protection for the contemporary refugee regime.
How should responsibilities for refugees be distributed? According to the UNHCR, at the end of 2014 there were 19.5 million refugees among a total of 59.5 million forcibly displaced persons worldwide.  Developing countries hosted 86 % of this refugee population (up from 70 % ten years previously.)  Lebanon (26 %) and Jordan (9.8 %) have the highest per capita ratios of refugees worldwide.  Is this a fair distribution of responsibilities? This question is not merely one of interest to political philosophers. Considerations of fairness have been much to the fore in the political rhetoric of debates concerning current flows of Syrian refugees into the European Union (although to put this into perspective, from the beginning of the crisis up to the end of 2015, the total number of asylum applications from Syrians in the European Union reached 681,713,  while in the same period the number of Syrian refugees in Turkey amounted to 2.18 million ). But at least one of the difficulties in this debate is that there is no agreement among states, globally or within the EU, concerning what would count as criteria of a fair distribution of responsibility for refugees.
The current EU crisis also illustrates a further question that is urgent in the contemporary context: what are the limits on state’s obligations to refugees? Is it, for example, sufficient to have done one’s fair share or, in the absence of established criteria, to have done what a good faith effort to work out one’s fair share required? Or do states that have done their fair share have an obligation to take up the slack consequent on others failing to do their fair share? If so, what limits are there on this slack-assuming obligation?
The central topic addressed in this essay is the relationship of duties of justice towards refugees and the distribution of responsibility for the protection of refugees (with particular focus on the questions of whether doing one’s fair share of refugee protection entails that one has done all that is required to discharge one’s duty of justice towards refugees). I address this topic in relation to the institutional context of the current international refugee regime. In the first part of the essay, I lay out the normatively salient features of the current refugee regime and the relationship between duties of justice to refugees and the fair distribution of refugee protection that it institutes. In the course of this discussion, I draw particular attention to the fact that it is an important feature of the current refugee regime that it does not limit the duty to protect refugees to doing one’s fair share of refugee protection. The second part of the essay considers whether this feature of the refugee regime is justifiable by addressing arguments concerning whether duties of justice are limited to doing one’s fair share of refugee protection. In particular, this section addresses the argument proposed by David Miller (2007, 2013, 2016) that there is no duty of justice to take up the slack that is produced by the non-compliance of other actors. It argues that Miller’s argument is invalid and that duties of justice towards refugees may require doing more than one’s fair share, but also that there is an obligation to seek the fairest arrangements compatible with effective refugee protection. In the final section, I consider the significance and difficulties of the issue of a fair distribution of refugee protection for the contemporary refugee regime in relation to an argument advanced by Matthew Gibney (2015) suggesting that while Gibney’s proposal has much to recommend it, it may underestimate the problems it is liable to confront.
In a just global order composed of autonomous states, each state would have both the capacity and disposition to secure the human rights of its citizens (and states would reciprocally protect the human rights of one another’s citizens where these individuals reside within a state that is not their own). This is not the world we inhabit. Some states lack the capacity to offer secure protection of the human rights of their citizens. Some states lack the disposition to do so. Some lack both the capacity and the disposition. We thus confront a condition of partial compliance with the duty of justice that states have to secure the human rights of their citizens.
Where a state disposed to secure its citizens’ human rights lacks the capacity to discharge this duty of justice, the international order of states has a duty of justice to support this state in building its capacity for securing the human rights of its citizens. Insofar as the international order of states in cooperation with the capacity-lacking state cannot secure the human rights of all of its citizens in situ, this may give rise to duties on other states to provide rights of (presumptively temporary) refuge until an adequate capacity for protecting human rights in the home state has been built. (This is particularly important in contexts in which domestic or transnational non-state actors exploit the state’s lack of capacity to threaten the human rights of some citizens/habitual residents or groups thereof.) These specific duties – to build capacity and, where needed, to provide refuge – are grounded in the general obligation of the international order of states (as a global structure of rule) to secure the human rights of all human beings (as those subject to its rule) in ways compatible with the structuring norms of this international order (Owen 2016). The latter point is significant. Thus, for example, except under conditions of necessity, state A cannot develop the capacity of state B to secure the human rights of its citizens by forcibly annexing state B against the will of its people even if this is the most efficient way of securing their human rights because to do so would be to breach the autonomy of state B (where this is a structuring norm of the order of rule grounded in the right to self-determination of peoples). 
Where a state is not disposed to secure the human rights of its citizens, the international order of states similarly has a duty of justice to act to guide the conduct of this state towards protecting the human rights of its citizens (and, if this state also lacks capacity, to help support the building of its capacity when the capacities in question cannot be used for human rights abuse or when the re-direction of its conduct provides grounds of confidence for the belief that these capacities will not be used in ways incompatible with the protection of human rights.) While the dispositional problem persists, the international order of states has a duty of justice towards those whose human rights are unprotected. In the case of those unable or who have good reasons to be unwilling to flee the state, this duty requires – as far as compatible with the basic norms of the international order of states – the provision of protection (for example, so-called ‘safe havens’) within the abusing state. In the case of those able and willing to flee the state, the duty requires granting them the right of asylum in other states where (we have well-founded confidence that) their human rights will be protected. Linking these two duties is a third that has particular salience in the context of the fact that, given the basic norms of the international order of states, the legitimate ability of the international order of states to protect those who are unable or unwilling to flee an abusing state may be extremely limited. This is the duty to offer proportionate  support, as far as compatible with the basic norms of the international order of states, to enhance the ability to leave of those who are otherwise unable to leave the state and, where possible, to address the reasons that may make others, who otherwise have good reason to flee, unwilling to do so (for example, the reasonable fear that they will lose assets of material and/or symbolic value such as the land that has been in, and partly defined, the family for generations). This set of specific duties are, like those concerning capacity-building, grounded in the general obligation on the international order of states to protect the human rights of all human beings.
It is against this general normative background that we can approach the current international refugee regime as the legal institution through which protection of those who are willing and able to flee state contexts in which at least some of their human rights are unprotected. The current definition of the refugee is provided by international refugee law which, strictly speaking, consists of ‘international and regional conventions, General Assembly resolutions and resolutions of various ad hoc groups with regional competence (e. g. the EEC Ad Hoc Group on Immigration), customary law and domestic legislation.’ (Tuitt 1996, p. 9, see also Goodwin-Gill and Adam 2007). However, the primary instruments of contemporary international refugee law are the 1951 Geneva Convention Relating to the Status of Refugees (hereafter ‘the 1951 Convention’) and the 1967 New York Protocol to the Convention Relating to the Status of Refugees (hereafter ‘the 1967 Protocol’) and, taken together, these instruments provide the basic definition of the refugee:
One who owing to a well-founded fear of being persecuted for reasons of race, religion, nationality [as belonging to a ‘people’], membership of a social group or political opinion is outside the country of his nationality [as membership of a state] and is unable or, owing to such fear, is unwilling to avail himself of the protection of that country; or who, not having a nationality and being outside the country of his habitual residence … is unable, or, owing to such fear, is unwilling to return to it.
(1951 Convention Article 1 A (2) & 1967 Protocol Article 1 (2))
This specification of the criteria of refugeehood has been much discussed (Shacknove 1985; Gibney 2004; Price 2009; Lister 2013; Owen 2016; Miller 2016) and there are regional instruments that adopt more expansive criteria – for example, in Africa, the Organization of African Unity’s 1969 Convention Governing the Specific Aspects of Refugee Problems in Africa and, in Latin America, the 1984 Cartagena Declaration on Refugees. For the purposes of my current argument, however, I will bracket these debates on the issue of criteria of refugeehood in order to focus on the question of the duties of justice to refugees conceived as persons characterized by unprotected human rights, who have good reason to leave their state of nationality (or habitual residence) in order to secure protection of their human rights, and who apply for protection of their human rights having reached the territory of another state. 
More immediately salient for my current concerns is another feature of the current regime, namely, that it is structured in terms of the grundnorm of non-refoulement. The duty of non-refoulement is a binding obligation on any state to which a claim to asylum is made not to return persons who, on the basis of an impartial process of adjudication, are found to satisfy the criteria of refugeehood to the state from which they have fled or to another state in which they would lack protection of their human rights. More formally, the duty of non-refoulement
encompasses any measure attributable to the State which could have the effect of returning an asylum seeker or refugee to the frontiers of territories where his or her life or freedom would be threatened, or where he or she is at risk of persecution, including interception, rejection at the frontier, or indirect refoulement.
An important implication of this principle is that the current regime places no restrictions on the numbers of refugees to whom a state owes a duty of non-refoulement. This does not entail that refugees have a right to asylum in the state to which they make application (a possibility advocated, discussed and rejected in the drafting of both Article 14 of the UNDHR and the 1951 Convention on Refugees). On the contrary, as long as the duty of non-refoulement is not breached, states can come to bilateral, multilateral or omnilateral arrangements with one another concerning the distributions of the presence of refugees and responsibilities for the costs of refugee protection. This point is important – even if, in practice, the state of application is still typically the state of asylum – because it points to the fact that responsibility for the protection of refugees is specified by the current regime in a way that is maximally consistent with respect for the autonomy of states. Under this current legal structure, it is up to states, jointly or collectively, to work out terms of reasonable cooperation for the fair distribution of refugees; if, or to the extent that, they fail to do so, the duty of non-refoulement entails that the default condition is that the state to which application is made bears the responsibility of ensuring the provision of asylum.
This normative structure of the contemporary refugee regime involves an important normative claim and constructs a practical dilemma. The normative claim is that discharging the duty to refugee protection takes priority over fairness of distribution of the responsibility for refugee protection. The practical dilemma is that, to the extent that refugees are perceived as burdens on the states that protect them and the costs of protecting refugees become significant, the current regime gives rise to a situation in which states, collectively, have an interest in effective and fair refugee protection, but states, individually, have an interest in minimizing their burden. In his path-breaking study of the ethics and politics of asylum from the standpoint of realistic political theory, Gibney makes the following highly pertinent point:
Above all else … the state is fundamentally an answer to the question of who is responsible to whom in the modern world: states are responsible to their own citizens. The survival of the state as an entity over time rests, moreover, on its ability to portray itself convincingly as an answer to such a question. As a consequence, the claims of outsiders are assessed by states, including liberal democratic ones, through a logic that deprecates the interests and needs of outsiders – a logic that is exceedingly sensitive to the potential damage to its own authority involved in forcing its citizens to incur costs for the sake of strangers.
(2004, p. 211)
This does not entail that states that recognize and respect human rights will not be concerned to act on their duty to offer protection to refugees, although this may be the case.  However, it does provide reasons to think, first, that even states that acknowledge their obligations to refugee protection are liable to be highly attuned to perceptions of unfairness in the distribution of responsibility and, second, that in the absence of agreed criteria for the fair distribution of responsibility, their perception of their fair share is open to being skewed by their interest in minimizing the burden placed on their own citizens, not least through the adoption of a ranking and weighting of the diverse factors involved in estimations of the relative burdens on states that favours their own interests. 
In the remainder of this article, I address the two issues thus raised. The next section asks whether the normative ordering of the current refugee regime (in which effective protection is prioritized over fair distribution) is justifiable. The final section considers the relationship between effective protection and fair distribution further in the light of the practical dilemma to which the current regime gives rise – and focuses particularly on the importance of fair distribution to the achievement of effective protection in this context.
The most sophisticated version of the claim that the duty of justice owed by a state to refugees is limited to doing its fair share has been advanced by David Miller (2007 and, more fully, 2016).  In National Responsibility and Global Justice, Miller argues thus:
Realistically … states have to be given considerable autonomy to decide how best to respond to particular asylum applications: beside the refugee’s own choice, they are entitled to consider the overall number of applications they face, the demands that temporary or long-term accommodation of refugees will place on existing citizens, and whether there exists any special link between the refugee and the host community … The best hope is that over time conventions will emerge that distribute responsibilities in such a way that refugees from particular places become the special responsibility of one state in particular (or a coalition of states). There can be no guarantee, however, that every bona fide refugee will find a state willing to take her in. The final judgment must rest with the members of the receiving state, who may decide that they have already done their fair share of refugee resettlement. 
(2007, pp. 226f.)
In the more recent work Strangers in our Midst (2016, pp. 83–93), Miller spells out the argument concerning refugees more fully and does so by reference to a more general argument he advances concerning justice, fairness and partial compliance under conditions in which (i) a collective duty of justice falls on multiple agents, (ii) it is clear what a fair distribution of responsibility involves, and (iii) where costs of taking up the slack generated by the non-compliance don’t cross the threshold of being too demanding or burdensome that would limit a state’s obligation (Miller 2013). This general argument is more restricted than the particular arguments concerning refugees because (iii) brackets question of demandingness that limit the obligations of states; it does so in order to focus solely on the issue of fairness.
A particularly significant feature of Miller’s argument (and an important reason to focus on his version of this type of argument) is that, in advancing his fair shares objection to there being a duty of justice to take up the slack, he is not arguing that there is no duty to do so. Rather he is drawing a distinction between duties of justice and humanitarian duties conceived, respectively, in terms of duties that can characteristically be enforced (in appropriate and proportionate ways) without wronging the agent (duties of justice) and duties that cannot legitimately be enforced (humanitarian duties). The political salience of this distinction is that, on Miller’s account, states are obliged to comply with duties of justice and hence have no obligation of justice to consider the preferences of their citizens concerning whether to comply.  By contrast, states do have a duty to consider the preferences of their citizens concerning whether, and to what extent, to comply with humanitarian duties. On this view, it may be the case that a state that chooses to do no more than its fair share despite the dire need of unprotected refugees is acting selfishly, but it does not act unjustly. What, then, grounds this fair shares argument and the distinction between duties of justice and humanitarian duties?
Miller’s argument can be reconstructed in two steps. The first step runs thus:
In standard cases, justice involves a claim made by one person against another, or by a person against an institution such as the state. In the present case, the obligation must attach to the whole group of potential contributors, who do not constitute an agent in the same sense: there is no collective decision to contribute or not to contribute, but rather each individual member decides whether to contribute his or her fair share, or more, or less. What we do have, on the other hand, is an agreed assignment of responsibility within the collective: by stipulation there is no dispute over what the fair share of each member amounts to. So the collective obligation to protect the rights of the potential victims does not simply hang in the air, but is translated into a series of individual obligations and these it seems clear to me are obligations of justice.
(2013, p. 215)
The second step is this:
The key argument here is that because the collective responsibility to avert injustice has been fairly distributed, ex hypothesi, by doing my fair share I have discharged my obligation, and the injustice that remains, because of partial compliance, is the responsibility of the non-compliers, and only theirs.
(2013, p. 217)
Two further points are salient. First, as already briefly noted, Miller acknowledges that although compliers have no duty of justice to do more than their fair share (that is, they could only justly be forced to this share), they do have humanitarian duties to take up the slack given the dire need of the victims of injustice who remain unprotected. Although the primary complaint of these victims must be directed at the non-compliers who bear responsibility in justice for their unprotected condition, they can have a secondary complaint against compliers who do not take up the slack generated by partial compliance (2013, pp. 221–5). Second, Miller recognizes that there is one special circumstance in which compliers would have a duty of justice to take up the slack resulting from non-compliance, namely, where compliers and non-compliers form a group proper such as a team who undertake to perform a task and where the non-compliance of some members does not entail that the remaining members have no obligation to take up the slack in order to perform the task in question (2013, pp. 219f.). This, he claims, is not the case in the context of states confronted with refugee flows.
We can begin to analyze this argument in the context of refugees by noting that the existence of refugees is already itself a product of partial compliance with the duty of justice of states to protect the human rights of their own citizens (and, by reciprocity, resident non-citizens). This matters for two reasons. The first is that, if Miller’s general argument is correct, then on a first interpretation of this argument it becomes difficult to see how he grounds a duty of justice to refugees on the part of justice-compliant states in the first place. The second is that, even if we reject this first interpretation, it indicates a problem with the assumptions involved in his articulation of the fair shares argument. Let us consider each in turn.
Let us note that, for Miller, human rights represent the moral bedrock of the contemporary international order and this grounds the duty of justice of states, collectively, to protect human rights. Under favorable circumstances in which every state has the capacity to protect the human rights of its own citizens, this collective duty of justice is (let us stipulate) fairly distributed through two norms: (1) the duty of justice of states, individually, (a) to protect the human rights of their citizens and (b) not to breach the human rights of non-citizens or undermine the capacity of other states to protect the human rights of their own citizens, and (2) the duty of justice of states, jointly, to ensure that no human being is stateless. Some states comply; others do not. However, on Miller’s general argument, since compliers have done their fair share, they owe no duty of justice to the refugees that arise from the actions of non-complying states but only humanitarian duties.
Suppose, alternatively, that we inhabit unfavorable circumstances in which some states lack the capacity to protect the human rights of their citizens. In these circumstances, the duty of justice of states, collectively, to protect human rights entails, additionally to the duties owed in favorable circumstances, a collective positive duty of justice to ensure that each and every state has the capacity to protect the human rights of its citizens. This collective duty is translated into fair shares of responsibility for each member of the collective that amount to individual obligations of justice. Some states comply and some do not. Compliers who also comply with their duty to protect the human rights of their own citizens have no duty of justice to refugees, given Miller’s general account, since refugees are either a product of states that have the capacity to protect the human rights of their citizens but choose not to do so or a product of states lacking the capacity to protect the human rights of their citizens, but states that comply in relation to both dimensions of the collective duty of justice have already done their fair share. Responsibilities of justice for refugees lie with the relevant type of non-complier according to the causes of a given refugee flow.
If salient, these reflections suggest only that Miller’s presumption that states owe duties of justice to refugees (2007, 2016) is hard to square with his own general argument (2013) concerning duties of justice and humanitarian duties when we acknowledge that the production of refugees, in contrast to the classic ‘child who falls into the pond’ type of case, is itself already a product of partial compliance with duties of justice. Miller’s general argument would then entail the view that justice-compliant states owe humanitarian duties, rather than duties of justice, to refugees.
In response to the first interpretation of his view that underpins these arguments, though, Miller might object that it misses the point. The issue that he is addressing, he may claim, is contexts in which, regardless of cause, some persons are in dire need and a plurality of other agents are so situated that they can help. In this case, refugees are in dire need and some states are situated such that they can help. His concern is simply with the question of whether, considered thus, states have any obligation in justice to do more than their fair share of refugee protection. In a moment we will grant this objection for the sake of argument but before we do, it is worth noting that, against the background of a broader concern with global justice and the protection of human rights, it isn’t straightforward to see what justifies considering the issue of fair shares of refugee protection independently of other work of human rights protection in which states engage (such as capacity-building). Why, for example, could a state that did its fair share of capacity-building, while other states don’t, not legitimately argue that this should affect the determination of fair shares of refugee protection? But let us set this issue aside and assume that we have reasons to treat duties of justice to refugees as a discrete issue and address Miller’s argument on this basis. In doing so, I will offer two arguments. The first accepts Miller’s framing of the issue in terms of singular ‘rescue’ situations such as the child falling into the pond. The second is advanced in a context of skepticism towards this framing.
We can introduce our discussion by noting Miller’s stipulation that what counts as any potential helpers’ fair share of responsibility is known prior to any decisions of compliance or non-compliance by particular states arising. This stipulation entails that a given actor’s fair share is specified by what justice would require of them, given their relative capacity to help, under conditions of full compliance. The attractiveness of this assumption is that it seems counter-intuitive to say that what properly counts as a state’s fair share of responsibility varies according to the degree of compliance or non-compliance of other actors. The problem is that, although it is reasonable to point to unfairness in contexts where an actor has to assume additional burdens due to the non-compliance of another actor with their duty of justice, the argument that an actor need only do their fair share as a matter of justice where the extent of the duty of justice is specified under the assumption of full compliance and this assumption does not hold true represents a prioritizing of intra-group fairness (horizontal equality) among potential helpers over effective protection over the human rights of those in need.
To draw out why this is a problem let us note that, in contexts of remedial responsibility, if an actor has a duty of justice towards another actor, this is a duty to act effectively to remedy the injustice (subject only to whatever limits are specified by a justifiable demandingness criterion). The duty of justice puts the agent under an obligation to reflect on how to act in order to act effectively (that is, successfully to remedy the injustice). Under some circumstances, this may entail that the agent considers whether their duty is best or necessarily discharged by coordinating their actions with other agents and, under such circumstances, this agent would certainly have a responsibility to reflect on whether the agents with whom they may coordinate their action are liable to comply or not with the demands of such coordination. Suppose, for example, that Fred owes a duty of justice to George and discharging this duty is presumptively best or necessarily accomplished by coordinating with another agent. Fred opts to coordinate with Alf, whose reputation for unreliability is notorious, rather than Bert, who is credited as being generally (if not perfectly) reliable. In this context, Fred cannot disavow responsibility for the consequent failure of his actions to discharge his duty of justice to George even if he performed actions that would have been sufficient for the scheme of cooperation to remedy George’s condition had Alf done his part, that is, even if Fred did his fair share of the work required. Fred’s duty to act effectively in remedying the injustice to which George is subject encompasses his duty to pay due attention to the choice of the partner, in terms of their reputation for reliability, with whom to coordinate his actions when seeking to discharge his duty of justice. It is thus not true that simply doing that what would be one’s fair share exhausts one’s remedial duty of justice.
When we turn to a group of agents who have a collective duty of justice to an actor or group of actors, the remedying collective similarly have a shared responsibility to act effectively to remedy the injustice. Under conditions in which we can either confidently expect full compliance on a voluntary basis or construct mechanisms that reliably enforce full compliance in legitimate ways, then the duty to act effectively can be discharged by each member of the group doing their fair share in terms of their relative capacity to help. In these conditions, the values of effectiveness and of fairness align – and this provides a compelling reason, where possible, to construct relationships and institutions through which full compliance can be confidently expected. However, where such conditions do not apply, the members of the group have a joint responsibility to reflect realistically on what is required to act effectively to discharge the collective duty of justice and must therefore take into account the probability that some actors may not comply and that others may need to take up the slack if effective remedy is to be achieved. This can be approached in two ways. In the first, there is an initial fair division of responsibility and, following non-compliance by some, a re-allocation of responsibility between those who have already done their fair share (and further iterations as required). In the second, the potential helpers deploy their shared knowledge of the likely conduct of each other and divide up responsibilities on the basis of this prudential assessment of what is thus necessary. Both of these approaches may require some to do more than what would be their fair share under conditions of full compliance and allow others to do less. In general, there is good reason to prefer the first approach that makes explicit the legitimate expectations on each actor and provides them with the opportunity to do their fair share. Or, put another way, it is important as a matter of respect to give people the opportunity not to comply in a context where they are expected to comply, rather than simply assuming (even if with good reason) that they won’t comply.  However, under urgent circumstances in which the costs of partial compliance to the refugees requiring protection are too high, for example, under conditions where acting effectively requires acting in concert and, hence, a high degree of confidence that the individual actors involved will act as required, then the second approach can also be justified.
Note that those who are required to do more have a justified complaint against those who are allowed to do less, and, in some circumstances, this might be expressed through the application of penalties or sanctions against those who do less than their fair share. (This can be the case even with the second approach insofar as an agent can justly be held responsible for the reputation that gives others good reason to believe that they will not do their fair share.) The key point here is that, in contexts of remedial responsibility, a duty of justice is a duty to act effectively to remedy the injustice and that this requirement of effectiveness takes priority over issues of fairness in the final distribution of responsibility among members of the collection of potential helpers. The injustice of granting priority to fairness over effectiveness is that in discharging its collective duty of justice, the responsibility for doing so fairly (securing horizontal equity) falls on this collective and hence the costs of its failure to distribute responsibility fairly cannot legitimately fall on those to whom the collective duty of justice is owed. To insist that justice is limited to doing one’s fair share is illegitimately to shift the cost of the failure to ensure full compliance, when this occurs, from the group who owes the collective duty onto the group to whom it is owed. Thus, in cases of collective duties of justice, there is an obligation to accommodate fairness as far as possible within the duty to act effectively but not at the cost of failing to act effectively. 
In making this argument, we are drawn to confront Miller’s identification of duties of justice with duties that would not wrong an agent if they were coercively enforced – and this identification is, it seems to me, central to his alignment of the extent of one’s duty of justice with doing one’s fair share since this is what would be the extent of one’s duty of justice under conditions in which compliance was enforced. Must we drop this identification in defending the claim that duties of justice may extend to doing more than one’s fair share?
Suppose that either a group of actors divide up responsibility fairly and, following the non-compliance of some, others have to take up the slack to ensure effective action or, given conditions of urgency and risk to those suffering the injustice in need of remedy, a realistic assessment of the actors who compose a collective leads to a division of responsibility that requires some states to do more than their fair share and allows others to do less. It is then proposed that coercive mechanisms are imposed that would penalize states that did not meet their final or assigned responsibilities. Would this wrong those states that are required to do more than their fair share?
If it is the case that the use of available coercive mechanisms would also have been capable of changing the conduct of states who have been allowed to do less than their fair share of responsibility in ways that would alter what counts as the fairest division of responsibility compatible with acting effectively to discharge the collective duty of justice, then this would wrong the states that have been required to do more than their fair share. It would so because they are being asked to do more than would be required by the fairest realistic scheme compatible with effective performance of the collective duty. Here the appropriate response would be a re-division of responsibilities in the light of the changed character of the scheme of cooperation. (The new scheme may still require some to do more than their fair share, but would require less additional burden than the pre-penalties scheme.)
But if this is not the case, if the proposed coercive mechanisms would not be capable of making a difference to the conduct of states who do less than their fair share, then it does not wrong the states required to do more than their fair share on the condition that the penalties for non-compliance with final or assigned responsibilities on those doing less than their fair share are proportionately greater than penalties on those doing more than their fair share. States doing more than their fair share are not wronged by this arrangement because if they perform less than their assigned responsibility, they have acted in ways that undermine what is necessary under current conditions for the effective performance of the remedying action, that is, they have breached their duty of justice.
However, while this argument may suffice to counter Miller’s argument on its own terms, it is also worth noting that we have some reason to be slightly skeptical of the analogy with singular (i. e., one-off) rescue cases of the kind that Miller (as well as other defenders and critics of fair share objections) take as exemplary for considering the issue. There are three normatively significant dis-analogies between the case of refugees in our current global condition and standard singular rescue cases. The first is simply that states have consented to the 1951 Refugee Convention and 1967 Protocol that prioritizes effective protection over fair distribution – and, in so consenting, acquire obligations to act on the basis of their commitment, namely, not to engage in refoulement or to prevent asylum seekers reaching their borders to make a claim to refuge. The second is that, unlike the standard case of the drowning child, refugee flows are an all too normal feature of our global political landscape. This is why we have an international refugee regime. The relevant analogy is not with the one-off case of a child (or group of children) drowning in a pond, but more with a neighborhood swimming spot (a river bend, a disused quarry, etc.) where unsupervised children regularly drown or risk drowning and where there is a collective obligation on the neighborhood community to ensure that supervision and rescue capabilities are reliably available. The third is that the same agents that are collectively responsible for remedying the injustice are also collectively responsible for the enforcement – or, more to the point, absence of enforcement – of the duties of justice in question. Duties of justice are, we have noted, duties such that their enforcement would not wrong the agents subject to them; however, if it is these same agents who are responsible for establishing the relevant mechanisms of enforcement and, hence, the generally reliable performance of fair shares, then they cannot reasonably claim that their obligations are untouched by the fact that, having not established such mechanisms, some do not fulfill their fair shares. Suppose our neighborhood community required all adults to attend lifesaver swimming classes and to serve as lifeguards on a rota so that three trained adults were always present to supervise and rescue at the local swimming spot when required – and it has the legitimate authority and power to impose this rule via sanctions over members of the community but chooses not to do so. In this context, the fact that, faced with a rescue situation involving three drowning children, Alf finds himself with Bert who chooses not to participate, while Charlie has not turned up at all, does not mean that Alf’s duty of justice is only to do his fair share as specified under the presumption of full compliance, rather he is obligated to rescue as many as he can without unreasonable risk to his own safety. The fact that Alf is part of a community of agents whose collective choices result in the failure to ensure Bert’s and Charlie’s participation in doing their fair share generates an obligation on him to do more than what would be his fair share under conditions of full compliance – and does so independently of the view that one takes of duties of justice in relation to one-off rescue cases with random potential helpers.
Reflecting on the arguments offered contra Miller’s fair shares view, both those accepting his framing of the issue and those rejecting it, we may ask whether these arguments fall victim to the charge that those who take on more than their fair share fail to treat those who do not as responsible moral agents – and, hence, that taking duties of justice as anything other than fair shares fails to treat non-compliers as responsible moral agents? In response, we may note that, as Stemplowska argues:
We need to avoid conflating two senses of responsibility: anticipating failure means that we are not seeing agents as responsible in the sense of well-behaved, but we could still recognize them as responsible in the sense of having the capacity to act responsibly and fulfill their duties.
(2015, p. 14)
Miller’s argument conflates just these two senses of ‘responsible moral agent’ as expectation and as accountability. In the first, it refers to someone who can be generally relied on to comply with duties of justice. In the second, it refers to someone who can be held accountable for failing to act on duties of justice. The relevant sense in respect of whether this account treats non-compliers as responsible moral agents is the latter, not the former, and it does treat non-compliers as responsible moral agents in this sense. It is, for example, this which would, under appropriate circumstances, justify sanctions against non-compliers and it is also this which requires that any schedule of penalties with respect to assigned responsibilities of the kind discussed above must differentiate between those doing less and those doing more than their fair share, penalizing the former more heavily than the latter. It is also the case that we have good reason to act on the conduct of responsible moral agents in the second sense to bring them to conform to the conduct of responsible moral agents in the first sense. But this is just to say that, absent compelling reasons to the contrary, we should aim to institute schemes of cooperation that reliably generate full compliance or as close to full compliance as is compatible with wider duties of justice. 
In this section, I have focused on the issue of whether duties of justice are limited to doing one’s fair share and argued that, although we have good reasons to seek to establish schemes of cooperation in which acting effectively and doing one’s fair share align, to the extent that such conditions do not obtain (or are not feasible) duties of justice may require doing more than one’s fair share when this is necessary effectively to secure refugee protection. If cogent, this argument vindicates the normative ordering of the current refugee regime that prioritizes protection of refugees over fairness of distribution of responsibility for refugees and, against the background of the default norm of non-refoulement, assigns responsibility to states, jointly or collectively, to work out arrangements for the distribution of responsibility for effective refugee protection. However, it also draws attention to the fact that states have a duty to come to arrangements that, as far as plausible, aim at ensuring a fair distribution of responsibilities (horizontal equality). In the final section of this essay, I will focus on the importance of this point.
Recall the practical dilemma constructed under the current regime in which, if the cost of refugee protection becomes significant, a collective action problem can arise in which although states, collectively, have a prudential interest in an effective refugee regime, they also have a prudential interest in minimizing their own contribution. In the absence of agreed criteria of, and arrangements for securing a reasonable degree of, fairness, this dynamic threatens to erode the willingness of states to engage in the provision of effective refugee protection. Thus Hathaway and Neve (1997) in motivating the last serious proposal for reform of the international refugee regime noted explicitly that the lack of fair apportionment of refugee protection was a key contributor to the reluctance of states to admit refugees for whom, in virtue of the principle of non-refoulement, they thereby bear sole responsibility for what often amounts to indefinite protection. Even among those states generally disposed to act justly in relation to refugees, the fact that they have little control over the numbers of refugees that they may be required to take is prone to lead to measures designed to control their potential liability for refugee protection by obstructing access to the territorial border: air carrier sanctions, the use of extraterritorial zones at ports and airports, interdiction at sea, ‘first safe country’ rules.  These considerations suggest that the aim of ensuring the provision of effective refugee protection requires developing criteria of, and arrangements for, the reasonably fair distribution of refugee protection in order to motivate even justice-disposed states to act to secure effective refugee protection.
The type of responsibility with which we are concerned is remedial responsibility in which, other things being equal, the responsibility to help is distributed in terms of the capacity to help. In the case of refugee flows, this implies that, other things being equal, those states that have the capacity to protect the human rights of refugees should do so according to their capacity. We can distinguish between two ways of thinking about capacity that are salient to thinking to duties of justice to refugees and to the fair distribution of responsibility for the protection of refugees:
The absolute capacity of a particular state: this refers to the total amount of refugee protection that a given state can provide subject to whatever limits on its obligations of justice can be justified (for example, that a state has obligations of justice to protect refugees unless and until this undermines its capacity to secure the human rights of its own citizens).
The relative capacity of any particular state: this refers to the amounts of refugee protection that particular states can provide at the same level of civic burdensomeness (specified in terms of the metric that is used for determining the justified limits of the obligation).
Note that the total capacity of states considered collectively is the sum of (1) across all states that are obliged to offer refugee protection and that, where recourse to full capacity is required, (2) will be given by the ratios of (1) across the relevant states.
The first point to note, against the background of these distinctions, is that both (1) and (2) are potentially scheme-variable, that is, what absolute or relative capacity a state has may vary depending on the scheme of refugee protection that is adopted. There are two reasons for this variability. The first is that it may depend on the degree and forms of cooperation between states that characterize a given scheme of refugee protection (for example, it might be the case that regional co-operation between states augments their capacity for refugee protection). The second is that the factors determining the capacity of a given state for refugee protection are dependent on what the scheme takes as the relevant distributandum or distributanda. Thus, for example:
Scheme A takes ‘refugees and the costs of protecting them in the territory of the admitting state’ as the relevant distributandum and will refer to such factors as, for example, population density, GDP and integrative ability in working out a fair distribution of responsibility between states.
Scheme B takes ‘refugees’ and ‘the costs of protecting them in a given territory’ as distinct distributanda that gives rise to two dimensions of fair distribution of responsibility, for presence and cost respectively, where different factors will apply to each dimension.
There is no reason to think that ‘absolute’ or ‘relative’ capacity will be invariant across these schemes (given an appropriate method of summing the two dimensions in scheme B). Since the normative rationale is that of remedying the condition of refugees, however, we can reasonably assume the duty of justice that states, collectively, owe to refugees entails that they have an obligation to adopt whatever scheme maximizes effective refugee protection. In making that determination, it is important to take into account the different interests of states as well as the interests of refugees; a point that plausibly guides us towards separating ‘presence’ and ‘cost’ dimension of refugee protection as Hathaway and Neve (1997) suggested.
If we bracket the issue of scheme-variability as a potential site of reasonable disagreement between states concerning capacity, there is still the question of the factors and the weighting of the factors salient for determinations of capacity under whatever scheme is held fixed. In an important recent essay, Gibney has argued that ‘there is a strong case for considering the integrative capabilities of states in determining shares’ (2015, p. 9). By ‘integrative capacity’, Gibney refers to a combination of two objective measures: GDP and population size (2015, pp. 9f.). His claim is this:
Real-world debates over the admittance of refugees almost always appeal to one of three considerations: total numbers [of refugees], GDP and population size … these are the very same standards that UNHCR sees as important enough to include in its reports. The lack of consensus over which standards to use should therefore not be exaggerated. Furthermore, all of these ways of measuring refugee burdens seem tell a similar story: Southern countries host an inordinate refugee share of the world’s refugee population. To be sure, the quotas of individual states change a bit depending on which specific standard we use. But they do not change much. States, I think, could not reasonably disagree to a proposal that balanced the three main standards – population, GDP and refugee population – together to determine their shares, particularly given the perversity of the current distribution.
(2015, p. 10)
Regrettably, I am much less sure than Gibney that this deals with the problem of reasonable disagreement for two reasons. First, while I agree that GDP is a sensible measure, so too is GDP per capita measured in terms of Parity of Purchasing Power (PPP). So, for example, The Netherlands and Malaysia have a similar GDP (measured in PPP terms), but the former has approaching twice the GDP (PPP) per capita of the latter. Prima facie, the higher GDP per capita is more able to absorb the costs of refugee protection in terms of civic burdensomeness. A similar issue arises in relation to population. While population size does matter in terms of the potential effects of refugee populations on the existing population of the state, population density may be a more important measure of integrative capacity. So, for example, the Netherlands and Niger have similar populations but the Netherlands has a population density of over 400 persons per square kilometer, whereas Niger’s has approximately 15 persons per square kilometer. Prima facie, this matters for their ability to absorb more population in terms of civic burdensomeness. Second, Gibney limits his criteria to population size and GDP by rejecting what he refers to as ‘subjective’ factors – but, as Gibney himself argued in earlier work (2004) and as Carens (2013, p. 215) also proposes, integrative capacity is also related to the immigration histories of states. These observations suggest that the question of which factors to consider in judging integrative capacity may be more open to disagreement that Gibney supposes.
The third point to note regarding distributions of remedial responsibility is that capacity is determinate of initial shares of responsibility only on the condition of ‘other things being equal’. This points to the fact that it also matters whether, for example, an agent with the capacity to help bears outcome responsibility  for the circumstances that generated the refugee flows or, even in the absence of outcome responsibility, whether they have benefitted from these unjust circumstances. It may also matter whether a state has important historical links with the country from which refugees are fleeing or stands in close proximity to the refugee-producing state (and thus facilitates return when this is justified). The varied dimensions of remedial responsibility and the difficulty that attends any fixed ranking of them  entail that there will be grounds for reasonable disagreement concerning the fair distribution of responsibility even among reasonable states committed to human rights protection and acting in good faith.
However, as Carens (pp. 222ff.) has argued, morality and self-interest are likely to pull in different directions in relation to the question of fair shares of refugee protection. State perceptions of fairness and of what scheme or weighting of factors should be adopted are likely to be skewed, deliberately or not, by their self-interest, where the scope for reasonable disagreement becomes a resource for the disguised expression of self-interest. Rather than adopt any particular scheme or weighting of factors, therefore, it may be more sensible to try to reach agreement on the construction of a fair and impartial process for assigning responsibilities in particular cases. This might take the forms of a fair process of political negotiation or judicial arbitration/adjudication (with an appeal process), perhaps supplemented by a free market mechanism that, following an initial allocation of quotas, allows the trading of ‘refugee tokens’. 
If it is indeed the case, as I have proposed, that the willingness of states to engage in effective refugee protection is generally conditional on the distribution of responsibility being tolerably fair, then the collective duty of justice of states to secure effective refugee protection entails a derivative duty to seek to establish such a fair process for assigning responsibilities.
In this essay, I have been concerned with the relationship between the duty of justice to protect refugees and the fair distribution of responsibility for refugee protection. I have argued that, from a normative standpoint, (a) the duty to provide effective refugee protection has priority over the fair distribution of refugee protection but that (b) states ought to aim at the fairest distribution that is realistically compatible with effective refugee protection. I have also argued that, from a prudential point of view, securing effective refugee protection is likely to depend on achieving tolerably fair distributions of refugee protection. Both of these arguments point to the need to develop international institutions that provide a fair process for assigning refugee responsibilities. In the light of continuing refugee crises, where these ‘crises’ are both crises of production (the generation of refugee flows) and crises of response (the failure of effective and fair coordinated state action), the absence of such institutions demonstrates both the significance of the political achievement involved in the creation of the current refugee regime (for all its limitations) and the very real political challenges involved in reforming this regime even under circumstances where such reform would be both rational and just.  This does not preclude us, however, from making the judgment that, with the plausible exceptions of Germany and Sweden, European states are acting unjustly – and that the EU as a polity is comprehensively failing to discharge its duty of justice towards Syrian (and other) refugees.
This article is, in part, another step in an ongoing argument with David Miller and I am grateful to David for his continuing willingness to engage in this argument. I owe particular thanks to Simon Caney, Matthew Gibney and Andy Mason as well as two initially anonymous referees (one later self-outed as Joe Carens) and the editors of the special issue for their very helpful comments and suggestions on an earlier draft.
Caney, S. (2005). ‘Cosmopolitan Justice, Responsibility, and Global Climate Change’, Leiden Journal of International Law, 18: 747–75. Search in Google Scholar
Carens, J. (2013). The Ethics of Immigration (Oxford: Oxford University Press). Search in Google Scholar
Cohen, L.J. (1981). ‘Who Is Starving Whom?’ Theoria 47 (2): 65–81. Search in Google Scholar
Feller, E., Türk V., and Nicholson, F. eds. (2003). Refugee Protection in International Law: UNHCR’s Global Consultations on International Protection (Cambridge: Cambridge University Press). Search in Google Scholar
Gibney, M. (2004). The Ethics and Politics of Asylum: Liberal Democracy and the Response to Refugees (Cambridge: Cambridge University Press, 2004). Search in Google Scholar
Gibney, M. (2006). ‘A Thousand Little Guantanamos: Western States and Measures to Prevent the Arrival of Refugees’, in K.E. Tunstall (ed.). Displacement, Asylum, Migration: The Oxford Amnesty Lectures 2004 (Oxford: Oxford University Press), pp. 139–69. Search in Google Scholar
Goodwin-Gill, G.S. and McAdam J. (2007). The Refugee in International Law (Oxford: Oxford University Press). Search in Google Scholar
Hathaway, J., and Neve, R.A. (1997). ‘Making International Refugee Law Relevant Again: A Proposal for Collectivized and Solution-Oriented Protection’, Harvard Human Rights Journal 10: 155–87. Search in Google Scholar
Karnein, A. (2014). ‘Putting Fairness in its Place: Why There is a Duty to Take Up the Slack’, Journal of Philosophy 111: 593–607. Search in Google Scholar
Kuosmanen, J. (2012). ‘Global Protection of the Right to Asylum and Partial Compliance’, Global Justice: Theory, Practice, Rhetoric 5: 46–55. Search in Google Scholar
Lister, M. (2013). ‘Who are Refugees?’ Law and Philosophy 32: 645–71. Search in Google Scholar
Miller, D. (2007). National Responsibility and Global Justice (Oxford: Oxford University Press). Search in Google Scholar
Miller, D. (2011). ‘David Owen on Global Justice, National Responsibility and Transnational Power: A Reply’, Review of International Studies 37: 229–34. Search in Google Scholar
Miller, D. (2013). ‘Taking Up the Slack? Responsibility and Justice in Situations of Partial Compliance’, in Justice for Earthlings (Cambridge: Cambridge University Press), pp. 206–27. Search in Google Scholar
Miller, D. (2016). Strangers in Our Midst (Cambridge, MA: Harvard University Press). Search in Google Scholar
Murphy, L. (2000). Moral Demands in Nonideal Theory (Oxford: Oxford University Press). Search in Google Scholar
Owen, D. (2010). ‘National Responsibility, Global Justice and Transnational Power’, Review of International Studies 36: 97–112. Search in Google Scholar
Owen, D. (2016). ‘In Loco Civitatis: On the Normative Structure of the International Refugee Regime’, in S. Fine and L. Ypi (eds.). Migration in Political Theory (Oxford: Oxford University Press), pp. 269–90. Search in Google Scholar
Price, M.E. (2009). Rethinking Asylum: History, Purpose, Limits (Cambridge: Cambridge University Press). Search in Google Scholar
Schuck, P.H. (1997). ‘Refugee Burden-Sharing: A Modest Proposal’, Yale Journal of International Law 22: 243–97. Search in Google Scholar
Shacknove, Andrew E. (1985). ‘Who is a Refugee?’ Ethics 95 (2): 274–84. Search in Google Scholar
Stemplowska, Z. (2015) ‘Doing More than One’s Fair Share’, CSSJ Working Papers Series, SJ035, http://www.politics.ox.ac.uk/materials/events/SJ035_Doing_More_than_Ones_Fair_Share.pdf Search in Google Scholar
Tuitt, P. (1996). False Images: The Law’s Construction of the Refugee (London: Pluto Press). Search in Google Scholar