Coercing Compliers to Do More Than One’s Fair Share

Is there a duty to do more than one’s fair share of solving collective problems? If there is, can those who do less than their fair share coerce others to do more? These questions arise urgently in relation to the problem of refugee protection. The fact that various states host refugees to a dramatically different extent is due to a range of factors but the most prominent one is that, on the whole, states devote their differential capabilities to the aim of not hosting them. When some states fail to host – or otherwise assist – as many refugees as they should, must other states do more than their fair share in order to solve the problem of providing refugees with an appropriate place to live? I examine what I call the conservative, moderate and responsive answers to this question. I defend the responsive view and argue that there is a duty to do more than one’s fair share here and that such a duty is always enforceable.


Introduction
Is there a duty to do more than one's fair share of solving collective problems? If there is, can those who do less than their fair share coerce others to do more? These questions arise urgently in the face of catastrophic climate change, global poverty and resistance to oppression. 1 They also arise, as a sizeable literature illustrates (Hoesch 2018;Miller 2016;Owen 2016Owen , 2018Stemplowska 2016), in relation to the problem of refugee protection in a world in which many states refuse to open their borders to them. I focus here on this last problem, addressing a debate that has recently unfolded in the pages of this journal between David Owen and Matthias Hoesch. 2 Although the absolute majority of people never leave their states, there are millions of people who would qualify as refugees under the official, narrow definition of who is a refugee, let alone any half-decent definition. 3 The UN estimates that there are 25.4 million refugees, 40 million internally displaced people, and 3.1 million asylum seekers (e.g. those 'awaiting a decision on their application for asylum'). 53% of the refugee population are children under 18 (UNHCR 2018, pp. 2-3). The majority of refugees who manage to leave their states and arrive in host states end up arriving in developing states: developing states do the bulk of hosting refugees. In fact, '[d]eveloping regions hosted 85 per cent of the world's refugees under UNHCR's mandate, about 16.9 million people. The least developed countries provided asylum to a growing proportion amounting to one-third of the global total (6.7 million refugees)' (UNHCR 2018). The fact that various states host refugees to a dramatically different extent is due to a range of factors but the most prominent one is that, on the whole, states -as represented by their governments -do not want to host very many refugees. They persistently devote their differential capabilities to the aim of not hosting them. Thus even when we are uncertain what -for any given state -would count as doing its fair share of hosting (or otherwise assisting) refugees, especially given their varying roles in creating them, we can still often be certain that a given state will fail to do what would qualify as its fair share under any reasonable interpretation of what it might involve. When these states fail, must other states do more than their fair share in order to solve the problem of providing refugees with an appropriate place to live?
One prominent answer, which I call the conservative view, holds that there is a duty to do one's faire share but no more than that. This view has been developed by L. Jonathan Cohen (1981), Liam Murphy (2000) and David Miller (2011 -initially in other contexts, though Miller himself went on to apply it explicitly to the case of refugees (Miller 2016, pp. 83-93). His own position is that there is no en-1 I do not mean to imply that all these problems are equivalent: in some cases, for example, only joint action can have the desired effect while in other cases individual action can be of use. I also do not mean to imply that the solutions to the problems need not be sensitive to who is causally or morally responsible for the problem arising. 2 Their papers are full of insight and detail that, unfortunately, I will not be able to comment on. Nor will I detail how their writings led me to develop my own views and I regret that I cannot do them justice here. I am also putting aside the key complication that the existence of refugees is primarily a result of wrongdoing that calls on different responses from wrongdoers than third parties. I do not think that putting this complication aside makes this analysis futile but it makes it radically partial. 3 For example, Miller (2016, p. 77) expands the definition to include all those who are fleeing on account of their human rights not being met. forceable duty to do more than one's fair share though there remains a humanitarian obligation to do more. 4 The opposing view, which I call the responsive view, holds that there is an enforceable duty to do one's fair share as well as -under some further conditions that are differently specified by different proponents of the responsive view -to take up slack when others fail to do their fair share. This view has been advanced in the context of refugee protection by, among others, David Owen (2016Owen ( , 2018 and myself (Stemplowska 2016). Many others, say consequentialists such as Arneson (2004) or Singer (1972Singer ( , 2002, would agree that there is a duty to take up slack, as would Cullity (2006) and Karnein (2014). 5 Recently in the pages of this journal, Hoesh has articulated what he has called a moderate view. According to this view, there is an enforceable duty to do one's fair share of aiding refugees, plus an enforceable duty to offer protection to all refugees over and above those who, as aptly summed up by Owen (2018, p. 179), '(i) ...are on [the relevant state's] territory or at its borders and (ii) who the state in question cannot effectively transfer to a country where the refugee will be safe and protected from non-refoulement that has not fulfilled its obligation to the extent to which the state in question has fulfilled its own.' This enforceable duty, however, is coupled with the permission 'to use border policies to reduce the number of asylum seekers that reach its territory to its fair share if those border policies do not coerce third states to carry more burden than this state itself does'.
Although the responsive and the moderate views differ on the existence of a general enforceable duty to take up slack, the specific iterations of each view can end up in agreement on specific instances of permissible coercion since each view allows that under some conditions it may be permissible to coerce states to do more than their fair share. The views can also end up in agreement on specific instances of impermissible coercion. Thus, for example, according to both Owen, a proponent of the responsive view, and Hoesch, a proponent of the moderate view, a state that does its fair share of aiding refugees has the right to 'resist' coercion to take up slack that would be meted out by international institutions' (Hoesch 2018, p. 172). 6 As I will argue in section 5 below, I disagree with this. My own view, which I defend below, is that the duty to take up the slack is always enforceable. 7 I begin in sections 2-4 by adding detail to the conservative, responsive and moderate views that are meant to guide us through the moral problem of refugee protection. Like Owen, I side with the responsive view. In section 5, I address more specifically the problem of permissible coercion in the face of slack taking.
First, however, let me explain and fix my terminology. 8 In what follows I will talk of helping refugees and refer to them as victims who are in dire need. I do not do it because I think that they are helpless -indeed anyone able to set off for safety in a world so hostile to refugees is far from helpless. I do it because I think they are entitled to help even by states that bear no responsibility for their plight. 9 I also imply that refugees are a burden that must be born and, given noncompliance, their presence is the slack that must be taken. By doing this I do not mean to suggest that they are not also a welcome addition to any country decent enough to host them or even that their presence is ever a net burden, only that there may be costs to some forms of accommodating them especially since refugees are entitled to more than merely to physical safety. I also focus on hosting refugees even though protecting them could take alternative forms, such as setting up decent temporary camps, helping refugee-creating states to solve their internal problems or problems with aggressors, or paying for other states to host refugees. Perhaps with the exception of the last policy -brought to the world's attention by a memorable phone exchange between Trump and Turnbull (The Guardian 2017) -hosting refugees in one's country is currently the most effective way of helping them at least in the short term. 10 The objectionable nature of current refugee camps is highlighted by Hoesch (2018, p. 173) who points out that '[t]hese camps are not only badly equipped with regard to health and nutrition; first and foremost, they do not provide the possibility to start a new life'.
'Duty to aid' refers to the (collective as well as the correlate individual) duty to assist someone in dire need whether or not it involves doing more than one's fair share. When one's aiding involves doing more than one's fair share, I refer to it sometimes as 'slack taking'. The agents who do less than their fair share of fulfilling duties to aid are 'noncompliers' or 'slackers,' while those who do at least their fair share are 'compliers'; if they do take up slack, they are also 'slack takers.' 'Slack taking' refers to the action of contributing at least some of the equivalent of a noncomplier's fair share. Slack taking involves substantive unfairness between the compliers and noncompliers, that is, the unfairness is comparative and concerns tangible burdens that people are asked to bear relative to one another.
Let me also adapt a stylised case that we owe to Hoesch (2018, pp. 171-2) to illustrate the problem at hand. Suppose that a state F -Fair -does its fair share of hosting the refugees, state S -Slacker -does nothing or very little, while the remaining host states (say H1, H2, H3, Hn) do less than their fair share (but more than Slacker). Suppose also that there are refugees who would like to join Fair. 11 Must Fair do more than its fair share? Can anyone permissibly coerce Fair to do so? I defend the responsive view and answer 'yes' to both questions below.

What's Wrong with the Conservative View
The conservative view, as developed by Miller (2011) and Murphy (2000) before him, assumes that at some time T1 the problem that needs a collective solution is assigned to the relevant collective (a group of persons, a group of states, etc.) and any duties that fall on individual members of the collective are derived as fair shares of what it would take to solve the problem (assuming it can be solved). From this time T1 onwards the duty of a member of the collective -be it a state or an individual -is simply to do one's fair share. If some members of the collective fail to do their own fair shares, this does not affect the duty of the remaining members of the collective vis-à-vis the problem that gave rise to the duties. In this sense, one's individual duty is not responsive to the changing circumstances (and for this reason alone the view can be labelled as conservative). Importantly, the collective itself need not be a pre-formed group; all that matters is that its members are all under a duty because they can help solve the problem in question and the problem is weighty enough to generate duties for those who did not create the problem in the first place (if they did, we are in the realm of corrective justice rather than the problem at hand).
The first insurmountable problem faced by the conservative view concerns the selection of the time T1. Any such point T1 is going to be impossible to identify, wildly implausible or arbitrary. It is impossible to identify if T1 is interpreted as a point in time -in the actual past or in some counterfactual trajectory -at which there is full compliance of all with their duties: if there is full compliance, then we can indeed assign duties to everyone to deal with a given problem in ways that preclude anyone being in position to complain that they are being asked to do more than their fair share. But, of course, it is impossible to identify such a point. First, it is impossible to know what our collective duties could possibly be at point T1 in the past with full compliance, let alone extrapolate their content to the present. Second, it is impossible to know what our problems and duties would be if, counterfactually, no one ever failed to perform their duties.
In addition to T1 being impossible to determine, searching for such a T1 is odd. As Owen (2016, p. 150) puts it, 'the existence of refugees is already itself a product of partial compliance': it is odd to search for a point where no one can complain that they have more than their fair share to do when the point arises because some people acted wrongly. In response, one could insist that duties to aid arise only in cases of natural disaster but never in cases of noncompliance by others, hence we can avoid the oddness. But this would merely take us into the widely implausible (Cullity 2006). Surely, no one decent thinks that innocent third parties could not be under a duty to aid children orphaned by murder rather than natural causes. As I put it elsewhere, '[a]ny plausible duty to aid must already accept that non-compliers's conduct can increase the burden that falls on compliers. Otherwise there could never be a duty to aid in the face of dire need generated by non-complying actions, such as killings, theft, aggression, or exploitation. But 'the duty to aid is not meant to be triggered only by dire need arising out of the operation of the natural world or innocent human conduct.' (Stemplowska 2016, p. 594).
Alternatively, the point T1 can be fixed in an arbitrary fashion. We could say that from this point onward, fair shares are determined by what it would take to comply in light of all the duties to aid those in need that fall on us. Although the arbitrariness of any such point is off putting, Murphy (2000, chap. 7) has offered a rationale for thinking that we may have no option but to pick it. The duty to aid, he has argued, is conceptually irreducibly collective and thus only derivatively an individual duty. That is, any given individual duty is a fair derivation from the collective duty that is fixed in relation to the original collective duty. Thus, unless we can identify some point at which the derivation can begin and the formula for the derivation then we cannot get to individual duties at all. Say we want to find out what a fair derivation to each individual is. Unless we know what the collective duty amounts to, we cannot do a fair division to get to the individual duties. On the assumption that we have no idea what the collective (problem and thus) duty would be in cases there was never any non-compliance, we assume instead that whatever needs fixating at time T1 provides us with the (arbitrary) starting point for the derivation.
Of course one natural response at this point would be that the derivation needs to be re-run each time there is need for duties to aid to be invoked. This is not Murphy's (or Miller's) idea, however. Rather, on the conservative view, once we picked T1, this is the original and only point for the derivation: from then on we cannot update the content of the individual duties.
As it happens, this inflexibility regarding a ban on updating the content of the individual duties as circumstances change provides us with a decisive argument against the conservative view. For if the individual duty is a once and for all fixed derivation of the collective duty, then the derivative individual duty remain fixed not only when the need for assistance grows but also when the need for it diminishes. Thus, as I argued elsewhere, crediting Victor Tadros with the key idea: ...[C]onsider a variant of the...pond scenarios [made famous by Peter Singer]. Assume that a fair share for 1 of 10 adults by a lake is to save 1 of 10 children drowning in it. Imagine that one person unilaterally saves all 10 children at no detectable cost to herself. But if, as Murphy holds, what one is under a duty to do does not depend on what others in the group do, the remaining 9 people seem to remain under a duty to save a child each. This should strike us as bizarre. It might be objected that the strange conclusion does not arise: the duty disappears because it is no longer possible to fulfill it (just as it would be had the children climbed out themselves). But this defense already concedes the main point: one's duty is not defined simply with reference to one's initial fair share but with reference to what remains to be done in light of the actions of others. (Stemplowska 2016, p. 595) It could be objected that it is not counter-intuitive to think that the duties of the remaining 9 would-be rescuers persist since they would still have to compensate the super-rescuer for his labour or loss. But suppose there are no compensation costs: what could the persisting duty amount to? Plus, in any case, any compensation would be owed to the super-rescuer rather than the victims, but the original duty was owed to the victims. Clearly, the actions of the super-rescuer simply extinguished those duties once and for all.
For those, as well as other reasons adduced by Owen (2016), Hoesch (2018), and others, such as Cullity (2006), the conservative view is mistaken and should be rejected.

Why the Responsive View?
On the responsive view, the duties individuals have to aid others are sensitive to whether noncompliance of others leaves more people in need of aid. It holds that instead of thinking of the individual duty as a fair derivation of the collective duty, we should think of it as a duty to do whatever it takes to solve the problem. As Garrett Cullity (2006, p. 75) has elegantly put it in his ground-breaking book, The grounds for thinking that beneficence requires me to ... [take up the slack and save the extra person left behind by the noncomplier] are plain. They are just the same as they would be in a situation in which I also could easily rescue one person: he desperately needs it, and I could easily help.
The fully just solution will need to be fair among those collectively under the duty but solving the problem takes lexical priority over ensuring the fair distribution of burdens if a fair distribution is incompatible with an effective solution. 12 This is, roughly, the view I take Owen to be defending in his paper in this journal. 13 Thus, for example, Owen (2016) writes that we should not prioritise intragroup fairness over effective protection of refugees (152) and holds that effectiveness takes priority over fairness (154). 14 It is also the view I defended elsewhere, objecting to the idea that '...the constraint for the derivation of individual duties be that of fairness rather than, in lexical priority, effectiveness, and fairness' (Stemplowska 2016, p. 596). Our negative duties clearly respond to the noncompliance of others since if I see you put poison in a glass I should not pass that glass to a thirsty person while there is no prohibition for me to pass the glass without poison to her. Our positive duties to aid should be no less responsive to the noncompliance of others.
That said, my own defence of the responsive view is limited to cases of aid in cases of dire need (so not all need that might trigger duties to aid). In cases of dire need, eliminating or minimizing it will trump our concern with inter-group fairness because the only defence against being put under a duty to eliminate dire need is that the burden is unreasonable. But inter-group unfairness alone would not render the 12 What counts as an effective solution here would, ideally, need more discussion than I have space for here. For example, it would be wrong to move refugees every 5 years even if this would facilitate fairness among the host states. 13 As he puts it there is 'an obligation to seek the fairest arrangements compatible with effective refugee protection' (Owen 2016, p. 143). 14 In Stemplowska (2016) I refer to the conflict as one of inter-group and intra-group fairness, holding that there is substantive unfairness if host states fail to help refugees. burden unreasonably heavy when otherwise it would be acceptable (Stemplowska 2016, pp. 596-7).
Owen offers two key defences of the responsive view. The first, I take it, is that those under a duty are required to be effective. I find this idea plausible when interpreted as a requirement to take reasonable steps to succeed. The second defence concerns cases when there is a 'community' such that we are collectively responsible for 'the scheme of rescue and the options for non-compliance within it' (Owen 2018, p. 182). One thought here is that in cases where the community exists agents have to think of who to coordinate with and how to coordinate. This is because there is a 'joint responsibility to reflect realistically on what is required to act effectively' (Owen 2016, p. 153). And as Owen (2016, p. 156; emphases in the original) also argues when it comes to refugees, we are facing not instances of one off, uncoordinated rescues one after another, but are under a collective duty to establish enforcement mechanisms. If someone is part of a collective that fails to enforce fair share taking on all, this generates 'an obligation on him to do more than what would be his fair share under conditions of full complianceand does so independently of the view that one takes of duties of justice in relation to one-off rescue cases with random potential helpers' (Owen 2016, p. 157; emphasis in the original).
I agree with Owen that all states should be seen as a community of states charged with the task of developing proper enforcement mechanisms dedicated to refugee protection. However, I still think that it is likely the presence of dire need alone that explains why states have a duty to do more than their fair share. The presence of a community that can coordinate enforcement matters in that it increases the capacity of the group to eliminate the need but does not change the fact that the reason to eliminate the need is grounded in the presence of the need rather than the presence of the community. Indeed, if need alone was not sufficient to ground the duty to aid, then any obligation to do so would have to be owed not to the victims but to the members of one's community.
I also worry that Owen's defence of the responsive view is too general. 15 I am not certain, for example, that the duty to take up slack follows simply from the fact that there is a duty of justice to deliver some outcome that falls on a collective. There is a collective duty (presumably of justice?) for the faculty to mark exams but I do not think -though I present this here as a mere assertion -that there is a duty to take up the slack. Similarly, I doubt that there is a duty to take up slack when tidying up after a primary school event even if the parents owe a duty to the school to leave it tidy.
But these are mere quibbles. Owen develops new arguments for justifying the duty to take up slack that emphasise the ongoing nature of the problem, and we are essentially in agreement here. Hoesch too offers an invaluable contribution to re-conceptualising the problem of slack taking, emphasising, with Owen, the way in which refugee protection is a problem that arises in circumstances when both the host states and the refugees can constantly, with greater or lesser difficulty, adjust strategies and respond to changing circumstances, which should affect what we morally ask of them. But I think the moderate view Hoesch offers still insulates compliers from slack taking too much.

What's Wrong with the Moderate View
Hoesch rejects the conservative view because he thinks -in line with international law -that it is impermissible to send away refugees who arrive at the state's border even if the state has already done its fair share (at least assuming the state has the capacity to accommodate them). The international principle of non-refoulement, which prohibits direct or indirect sending of refugees to unsafe places, takes priority over the consideration of fairness. To that extent Hoesch's moderate view accords with the responsive view and disagrees with the conservative one (though, it is worth highlighting again, that proponents of the conservative view have explored other strategies to avoid the conclusion that one is morally off the hook when refugees knock at our door but we have already done our fair share). 16 However, just like the conservative view, the moderate view resists the conclusion that what the compliers must do should be directly sensitive to how much noncompliance there is by others with whom they share the collective duty to aid. This is because, according to the moderate view, while a state must take in refugees who are now knocking on its door, if the state has fulfilled its fair share of refugee protection, it can also take steps to stop refugees from knocking on its door. There is no duty to take in more refugees than one's fair share just as long as the refugees are not knocking on one's door. As Hoesch (2018, p. 170) puts it, states who have done their fair share 'may try to prevent refugees from reaching their borders'.
Does this permission to stop people from reaching one's border extend to cover cases of all refugees? Consider two possibilities. Suppose first, counterfactually, that all those in need of protection find it in some reasonably safe host states, but the distribution of the refugees between the states remains deeply unfair. Some (hitherto) refugees would prefer to reach other host states than those they are currently in. There is no doubt that in such a case, according to Hoesch, a state that has done its fair share has no extra duty to take those willing to change their host states.
But what does the moderate view imply in a different case: one in which those who wish to reach the shores of the relevant state are not yet in a place of safety (e.g. remain refugees)? Is the permission extended to states who do their fair share of aiding to prevent refugees from reaching their borders meant to apply to such a case? After all, to be a refugee one must still be in some sense in peril and the moderate view is meant to inform states policies towards refugees.
How to interpret the view? On the other hand, Hoesch is clearly focused in the relevant part of his article on the problem of transfer of refugees between various host states, which suggests that perhaps the permission to stop refugees from reaching one's borders is meant to apply only to cases when no one is in peril any more. Moreover, as Owen (2016, p. 145) observes, there is, under the international order, the duty 'to enhance the ability to leave of those who are otherwise unable to leave' and would qualify as refugees, and who cannot or won't be helped at home. Stopping refugees from reaching one's border, including not offering visas on application to the embassy, is the opposite of enhancing their ability to leave.
However, even if the permission is meant to cover only instances of stopping refugees from swapping host states, it is unclear that the moderate view, as currently on offer, has the resources to contain the permission to only such scenarios. After all, if a state has no duty to do more than one's fair share of helping except for those who knock on its doors, then it would appear that it has no duty to help anyone reach its doors and can erect barriers for people to reach them. Granted, no state would be permitted to trap refugees in the unsafe state but, for Hoesch (2018, p. 172), closing access to one's own border is not the same as trapping someone elsewhere just as long as they can reach other borders.
To see the implications of the moderate view interpreted in this harsher way, consider an analogy. Suppose there are 10 people drowning in a lake and 10 rescuers such that the fair share for each rescuer is to recue one person. Suppose also that the lake is demarcated into 10 zones, one for each rescuer, such that whenever people reach one of the zones, the rescuer assigned to the zone has a duty to rescue them, no matter how many that is. On the moderate view, a rescuer who has already rescued one person is permitted to fence off his rescue zone, preventing people from reaching it just as long as others have not fenced off their zones, even if he knows that the remaining nine rescuers will not in fact rescue anyone.
In my view, this is straightforwardly unacceptable. It also has the consequence that it, in fact, diminishes what would count as doing one's fair share, even if a given state at no point shirks in its fair share of duties. Why? Because limiting people's options to reach safety discourages them, makes it harder for them and may even stop them from trying to reach it, thereby reducing the overall number of migrating refugees. But of course a reduction in the number of migrating refugees is not the same as a reduction in the number of people who are in peril and in need of safety. In Europe, we are of course familiar with these arguments in the context of cross-Mediterranean migration. It is sometimes argued that rescuing people who travel on unsafe boats will only encourage others to do the same. This is true. Whether such an encouragement is regrettable, however, should depend on whether in its absence people stay safely at home or simply endure similar or worse risk in the course of their lives without even trying to cross the sea.
But even putting aside the harsh version of the moderate view, which perhaps Hoesch would himself reject, we may wonder on what grounds the moderate view rejects the existence of the duty to take up slack. After all, as Hoesch points out, the conservative view is mistaken, so other grounds are needed. On my reading of the view, the main argument relates to the impermissibility of coercing those who do their fair share to take up slack by international organisations or the majority of the world states. 17 I address this in the section that follows.

Coercive Slack Taking
When is it permissible to coerce states to take in refugees? According to Hoesch (2018, p. 173), there is a duty not to coerce others to do more than one does oneself. This means that a state that does less than one's fair share cannot coerce another state to even do that state's fair share. By contrast, if the coercer does at least his fair share (or, perhaps, simply does more than the state that it wants to coerce), then the coercer has a right of 'self-defence' to coerce that state (Hoesch 2018, p. 172). I am not sure why we should see it as a right of 'self-defence' but let me put this to one side and focus on cases where the moderate view finds coercion impermissible. It will include instances of potential coercion by international institutions since, as Hoesch points out, international institutions contain non-complying states and thus coercion by international institutions would fall foul of the postulated duty not to coerce others to do more than one does oneself. As he puts it: 'Why should the collective, substantially constituted by those selfish states, have a right to coerce [a state that does its fair share] to unburden [a state that does more than its fair share]?' (Hoesch 2018, p. 172).
One possible thought here might be that states that do less than their fair share have no right to coerce others to take up slack since they could instead simply eliminate the problem by doing their fair share themselves. This is intuitive, though it also has far reaching implications: on this view it looks like the international collective cannot even coerce states to do their fair share since it is itself composed of many who do basically nothing or, worse, create the problem that needs fixing in the first place.
I dispute below the idea that the absence of the right to coerce makes coercion impermissible but first let me pause to outline also Owen's take on permissible coercion as set out in his original article. There Owen (2016, p. 155) considers whether coercing states to admit refugees would wrong them. His answer is that it depends. Coercion does not wrong if it can either (1) be used to increase the initial compliance of states with their duty to do their fair share (without sacrificing effectiveness of our efforts to give refugees their due), or, in case we cannot do that, if (2) coercion is used to force states to take up slack but the penalties imposed on them are proportionally lower than those imposed on the non-compliers.
The rationale for the first condition seems to be clear: we can coerce states to meet their initial, enforceable duties. This is of course the key claim of the responsive view and the key way in which, in my view, Owen's argument is correct. The 17 This, of course, would not eliminate the possibility of their being an unenforceable duty to take up slack. As I understand it, Hoesch's (2018) main argument against the existence of (even an unenforceable) duty to take up slack is that doing so, would not eliminate inequality between host states. But the mathematical example he gives (171) does not show that in some instances, taking up slack might reduce inequality. In addition, equality is not the only value at stake, not even for egalitarians. For criticism of Hoesch's argument here see Owen (2018, p. 183). rationale for the second (disjoint) condition, however, seems less obvious to me and I worry that it would rule out as impermissible too many cases of possible enforcement. Coercion is here clearly assumed to include both sheer force as well as the threat of penalties. And, notice, that the extent of penalties threatened in case of wrongdoing usually track not only the magnitude of the wrong we wish to prevent (and/or punish) but also the risk of the non-complier being caught (the lower the risk the higher the penalty in order to bolster its disincentive effect) as well as the pain the penalty would actually cause (also in order to preserve the disincentive effects on penalties used against targets of varying riches and capacities). But if penalties do not simply track the magnitude of the transgression, then it may not be essential to set the magnitude of penalties imposed to motivate slack taking at a level lower than the penalties imposed to motivate the initial compliance. A further reason to be sceptical of (2) is that the condition may in fact prevent penalties from tracking the magnitude of the transgression alone. For example, suppose that each of the initial non-compliers shirk in their duties only to a small extent as measured by what they are required to accomplish (though with potentially tragic consequences) in that they leave, say, 'only' 10,000 people each in peril. At the same time the slack taking that we could enforce on the initial complier, given his continuing capacity to aid, would force him to aid 100,000 people. It is not implausible to think that the coercive measures used against the initial 10 non-compliers should be lower than those used against the initial-complier-and-slack-taker who now has a more urgent task to perform. I should add, however, that Owen states the second condition only in passing and, as it is not essential to his core argument, it could be easily revised without sacrificing anything of importance in his overall position.
In his later paper, Owen (2018) responds to and agrees with Hoesch regarding the permissible use of coercion by international institutions (though, again, it seems to me that Owen is not committed to this view by anything he says elsewhere in his paper). It is here that I disagree with both theorists. 18 To explain my position let me revisit the scenario set out in the Introduction. On the responsive view, Fair has an enforceable duty to take up slack. But does Slacker have the right to coerce Fair to take up the slack? Let me put aside the easier cases, in which Slacker is no longer able to comply at this later point even if it tried (perhaps it had destroyed its capacity to take in refugees) or in which Fair itself could coerce Slacker to do Slacker's fair share. Suppose, instead, that Slacker has the capacity either to protect the refugees itself or force Fair to do so and these two options are the only options that would deliver refugee protection.
Clearly given its option set, Slacker is never in a position in which it can appear necessary for Slacker to force Fair to take up slack in order to protect the refugees. At the same time, the responsive view holds that, if Slacker will not comply, then Fair has an enforceable (remedial) duty to take up the slack. And although Fair cannot permissibly resist coercion from Slacker, Fair is wronged by it. Fair has a right against Slacker that Slacker does not coerce Fair, though, in this case, this right itself cannot be permissibly enforced given Fair's duties towards refugees. And, after all, we are all familiar with rights that cannot be permissibly enforced in a given situation: this is what I have against you when in order to prevent you from stealing my car, I'd need to kill you or shoot a bystander in the leg. The fact that I cannot resist you stealing my car under the circumstances would not mean that you did not wrong me.
So in the case of coercion by Slacker, Fair retains two complaints and two claims to compensation against Slacker: first, for needing to take up the slack and, second, for being coerced to do so by Slacker. But this does not mean that it can resist coercion: it has a duty towards refugees not to resist. We should therefore agree with Hoesch and Owen that Slacker, perhaps including some international organisations that contain Slacker, 19 does not have a right to coerce states such as F to take up slack in that Fair does not have a duty to Slacker to accept the coercion and Slacker wrongs Fair in coercing it. But, nonetheless, Slacker does have a conditional permission to coerce Fair on behalf of the refugees when S itself will not help them and Fair has a duty towards the refugees not to resist.

Complications
Sadly, none of what I say here actually offers a practical solution to the problem posed by states whose duty it is to aid refugees but refuse to do so. It also does not advance our thinking on what might count as the fair share we would expect states to do. Owen (2016, p. 162) himself offers an illuminating brief discussion of the difficulties of agreeing on what might count as a fair share and does so since 'securing effective refugee protection is likely to depend on achieving tolerably fair distributions of refugee protection'. He ends up suggesting that we may try to consider as fair shares what is agreed as such by an impartial process. This is a sensible suggestion though it is doubtful whether any such process will be effective in the face of wide disagreement by parties to the process on the substantive question of fair shares. Thus I have one further proposal to add to the pot: a state that opens up its borders to refugees to such an extent that it makes it rational for refugees (even if not all migrants) to abandon alternative routes of travel over there -without also increasing penalties and frequency of capture for those who travel without the required documentation -is on its way towards meeting its fair share (It is 'on its way' since some states will 'benefit' more than others from their distant or otherwise hard to reach location from where refugees originate in terms of how likely refugees are to reach them.). That said, as my argument makes clear, answering what might count as a state's fair share is less important than we might suppose since in the presence of slacking by some states, other states in any case must do more than their fair share.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.