The Openness-Rights Trade-off in Labour Migration, Claims to Membership, and Justice

This paper looks at a recent challenge to the liberal inclusivist view that everyone on the state’s territory should have a path to citizenship. Economists have argued that giving immigrants an inferior legal status would persuade wealthy countries to admit more, with beneficial consequences for global justice. Whilst this trade-off might seem appealing from the impersonal perspective of the policymaker it generates incoherence from the perpective of the collective of democratic citizens, since it requires them to treat their own unjust attitudes as an objective constraint. The paper also rejects the idea that a voluntary choice to migrate can be taken as consent to an inferior status.

In an early contribution to the philosophical debate around migration and justice, Michael Walzer argued that temporary labour migration is anomalous and that such migrants, if their stay extends beyond some bare minimum, should be put on a path to full citizenship (Walzer 1983, pp. 56-61). What is intolerable, according to this liberal inclusivist view, 1 is that some residents on the territory of the state should be consigned to a long-term or even permanently inferior status. In the recent past, some, particularly economists, have argued that such inferior statuses for labour migrants are not only permissible but perhaps even morally required to reduce global poverty and inequality.
I set out such arguments and respond to them, distinguishing between empirical and normative issues. I argue that the argument that we should offer some migrants a reduced set of rights to increase migration flows in the interests of reducing global poverty depends crucially on taking a detached perspective and seeing citizens and their legislation merely as one of the causal factors in play as we try to achieve an impersonally best outcome. This detached perspective contrasts with one that sees the demos of a democratic state as a collective agent with a responsibility to behave justly. When we and our fellow citizens deliberate and decide what our migration and citizenship policies should be, we cannot treat our own reluctance to act justly as an objective external constraint on our own choices. If migrants have moral claims to certain legal rights, including rights to citizenship, or that the state not knowingly expose them to domination or exploitation, we cannot simply set those claims aside in the name of maximizing the good or promoting a globally just distributive outcome. 2 Though I argue that such inferior membership statuses should be rejected as unjust, this does not necessarily apply to all limitations of rights for temporary labour migrants. If states have the right to control migration then some restrictions, such as time-limited visas or reduced access to welfare schemes cannot be ruled out as impermissible. 3 I also reject arguments that migrants themselves can be thought to have implicitly consented to an inferior status.

The Argument for a Rights-Openness Trade-off
The argument that we should limit the rights of migrants in order to tackle global poverty and inequality has been advanced by people including Martin Ruhs (2013), Eric Posner and E. Glen Weyl (2014;, Daniel Bell and Nicola Piper (2006), Dani Rodrik (2011), Lant Pritchett (2006), and Branko Milanovic (2016). The argument has empirical and normative components, but what I shall be most interested in here is the normative side. I shall mainly concentrate on Branko Milanovic's version. This is because it he makes a clearly articulated and general case which sits in the middle of a spectrum of variants of the argument. Because he is willing to challenge migrants' long-term access to political rights in order to boost numbers, Milanovic offers us a good example where there is a tension between global distributive goals and the denial of rights that are both widely supported and plausibly a requirement of justice.
In his book Global Inequality, Milanovic (2016) argues that the structure of inequality in the world at the moment is such that inequality between countries exceeds inequality within countries. As he puts it: The global Gini value of slightly under 70 is significantly greater than the national Gini value in even the most unequal countries in the world, such as South Africa and Colombia. (2016, p. 132) Location, then, is currently a more important source of inequality in the world than class. Where location is significant as a source on inequality in this way, citizens of wealthy countries earn what Milanovic calls a Bcitizenship premium^, which he describes as being in essence a rent, or if we use the terminology introduced by John Roemer in his Equality of Opportunity (2000) it is an Bexogenous circumstance^(as is the citizenship penalty) that is independent of a person's individual effort and their episodic (that is, not birth-related) luck. (2016, p. 132) The basic point is clear: the world economy is structured such that some people enjoy citizenship statuses that are much more economically valuable than others. This is a benefit or a penalty (depending on the case) that does not arise from the choices or efforts of the holder but wholly from their unchosen circumstances. 4 Not only does the citizenship premium give this advantage to those lucky enough to be born in wealthy states, the existing migration system also reinforces the benefits that accrue to other unchosen aspects of individual endowments. People Bwho were not lucky enough to have been born in a rich country but have exceptional abilities or wealth^(2013, p. 136) can relocate to wealthy countriesperhaps under BAustralian-style points-based systems^augmenting their own advantage and that of their new country to the detriment of the lessadvantaged who are locked out of the rich world by exclusionary migration regimes.
So what is to be done? Milanovic points to the uncomfortable example of regimes like the Gulf states which import high numbers of temporary migrants to whom they deny many rights and any path to permanent membership. As he puts it … it could be argued that the Gulf countries by welcoming foreign workers en masse are actually contributing effectively to the reduction in world poverty and world inequality. … The less harsh and yet still discriminatory treatment of migrants in the rich countries could have even more beneficial effects. But to take that step, one would have to accept what seems like a huge shift in policy: discriminatory treatment of migrants in the recipient countries, and de jure introduction of two or three levels of Bcitizenship^rights, at least for a while. (2016, p.151) He notes that this policy offends against conventional norms under which Bpeople in rich countries and their governments are very concerned with providing (at least, legally) equal treatment to all people living within the country's borders.^(2016, p. 150) He argues that it is inconsistent to be so concerned about the condition of workers inside a state but to be largely indifferent to their treatment in other countries. He also argues that this assertion of a norm of equality is contradicted by the de facto toleration of large numbers of irregular migrants in many countries: Many of these scenarios are now happening informally, as in the case of some 10 million undocumented immigrants in the United States who, because of their unsettled status, have to accept lower paying jobs. But such discrimination is not codified. In the eyes of many people, it therefore does not exist. (2016, p. 153) In any case, there is a further issue, and that is voluntary acceptance of lower status in return for economic benefits. As Milanovic puts it: … we can be quite sure that migrants would consider mild discrimination or unevenness in treatment in recipient countries to be preferable to remaining in their countries of origin by looking at their revealed preference …: their very willingness to migrate reveals their belief that migration would increase their welfare. (2016, p. 153) So not only does such an assignment of inferior status lead to a better global distributive outcome, it is also potentially made legitimate by the consent of those who get the subordinate status. 5 I return to the question of consent below.
By contrast with Milanovic, Martin Ruhs provides us with an example of a weaker version of the openness-rights trade-off in his book The Price of Rights. 6 Ruhs confines his attention to temporary labour migrants and argues that states are justified in restricting rights to family reunion, rights to some means-tested benefits and rights to seek employment beyond a specific sector. The first two restrictions are justifiable only when needed to protect the state from a net fiscal loss for the duration of the temporary worker programme; the second to ensure that the programme achieves its purpose of filling sectoral labour shortages. Beyond this, Ruhs argues that such migrants should have the full range of legal rights afforded to citizens, except for the right to vote, and that those who stay beyond 4 years should be entitled to apply for permanent residence and be on a path leading to naturalization. At the other extreme from Ruhs, and Weyl (2014, 2018) offer a version of the argument that is needlessly rhetorical and provocative and so might be thought to present a less challenging target for critique.

Empirical and Normative Preliminaries
Milanovic's argument has empirical and normative elements. One line of criticism would be to challenge its empirical basis. For example, the claim that citizens of advanced liberal democracies would be more open to mass migration to poor countries if only the immigrants were accorded a lower formal status might not be true. As it is, the claim that there is a causal link between limiting the rights of migrant workers and greater openness is based on noticing a correlation between the two which might not be genuinely indicative of causality and which might not generalise across different societies. For example, if in some advanced capitalist countries citizen hostility to migration is predominantly grounded in worries that inward immigration could reduce wages, it is hard to see why such hostility would disappear just because the new competitors have a vulnerable status that exposes them to further possibilities of competitive exploitation. In the UK, for example, nativist complaints about EU migrants have often focused on their willingness to work harder for less. Where citizen hostility to immigration is based on other factors, such as concerns about costs to public and welfare services, then Milanovic's empirical case be might more solidly based.
Although disagreements about facts can take on great moral significance and everyone, including philosophers and political theorists, has an interest in finding out the truth about them, my focus here is on normative principles. For the purposes of this paper, then, I shall take the facts to be as Milanovicalong with other partisans of the trade-offspecifies them. I accept for the sake of the argument that: (A) inequality between countries is greater than inequality within countries; (B) there is a large citizenship premium; (C) migration from poor countries to rich ones is a highly-efficient way of ameliorating global poverty and inequality; (D) a big obstacle to such migration is the attitudes of electorates in wealthy liberal democracies; and (E) there would be more poverty and inequality-reducing migration if wealthy countries were to restrict labour migrants to inferior political statuses. 7 We can ask whether, given these assumptions, it is morally permissible or even required to introduce such inferior statuses, whether some rights can permissibly be sacrificed and others not, and whether the consent of immigrants themselves makes a difference. For the purposes of his paper there is no need to be too precise about the distributive goal that is served by the openness-rights trade-off. Milanovic's own presentations suggests an affinity with luck egalitarianism and indeed with a view that gives luck egalitarianism global scope. 8 Fortunately, Milanovic does not need anything as controversial as a strong egalitarian premise to make a version of the trade-off argument since there is almost universal convergence around the idea that severe poverty is bad, and ought to be corrected. So, we should simply stipulate the causal fact that increased migration from poor countries to wealthy ones would reduce severe global poverty and the moral facts that the world would be better and more just if severe global poverty were reduced.

Challenging the Trade-off Argument
We can now address the normative issues posed by the alleged trade-off between, on the one hand sacrificing migrant's rights and according them an inferior political status and, on the other, promoting good global distributive outcomes. The central difficulty I see arises from the perspective from which the proposal is formulated and advanced to its intended audience. Milanovic and others who favour the openness-rights trade-off propose that liberal democracies water down or abandon one of their core normative commitments (status equality) for the sake of a cosmopolitan distributive goal (reducing severe poverty globally). But we need to ask why this incompatibility between status equality and openness exists. The answer, by assumption, is that voters cannot be persuaded to pursue the beneficial (and putatively more just) outcome without dropping the equal status commitment. To formulate the issue like this is to adopt the standpoint of the technocratic Bpolicymaker^concerned about producing the best outcome given realistic constraints. Citizens and their views are seen as obstacles that need to be evaded or manipulated if the best (or second-best) moral result is to be had.
This perspective is, however, not the only one available. This is because the policy is also a choice made by the electorates of liberal democracies or their representatives. Framed like this, it is a proposal about what we should do, and indeed what we may permissibly do. If a collective subject exists and is subject to moral constraintsincluding justice and individual rightsand is a suitable object of blame and responsibility, then it seems problematic for it to take its own attitudes (or those of its members) as a constraint in the manner that the trade-off argument does. What we have are some requirements of justice, namely the individual rights of non-citizen persons on the territory of the state, and a distributive goal. The bearers of at least some of those individual rights are entitled to have them respected and that means that othersincluding relevant collectives of personssee those rights as side-constraints on permissible action, particularly when the action in question can be undertaken whilst also respecting those rights. Although there may be cases of emergency or rescue where individual rights have to be overridden in order to avoid great harms, those cases do not plausibly include ones where the Bnecessity^for overriding rights is merely an artefact of the attitudes of the rescuer rather than being a more objective constraint. In the present case, the democratic public could if it chose realize both the valuable goal (relieving world poverty) and simultaneously admit immigrant workers on terms that respect the rights to which they are entitled, the problem is that it chooses not to. We should not then represent that choice as being morally required. 9 The point about moral choices and unjust attitudes as a constraint may be clearer and more intuitively obvious for cases involving the moral deliberations of a single natural person. We can then address the issue of whether the moral lessons of the single-person case generalize to collectives like democratic electorates. Consider, then, an example of individual moral agency with a valuable goal involving a positive duty of rescue (plausibly analogous to the goal of relieving global poverty) and a difficulty involving the unwillingness of the duty-bearer to act that arises from the unjust attitudes of that same duty-bearer: It is a cold and stormy night and Misanthrope (a misanthrope) is sitting cosily by the fire in his solidly built castle. Meanwhile, Hiker (a hiker who has accidentally wandered from a safe route) is in the cold outside and will perish from hunger and hypothermia if Misanthrope does not admit her to the warmth of the castle. Misanthrope hears the knocking and Hiker's plaintive cries for help at the castle gate and deliberates about what to do. Misanthrope has a stringent positive duty to rescue Hiker. Though Misanthrope's beliefs about whether he has such a duty are irrelevant to whether he has one, we can assume that even his determined and motivated study of the works of Ayn Rand has failed to dislodge his belief that he has a duty to rescue Hiker. So he has a duty and he knows that he has that duty. Though he is aware of his duty, Misanthrope is not always able to get himself to act as it requires and has a keen insight into his own psychology. Although he could simply run downstairs, open the door, and admit Hiker, Misanthrope knows that he can only motivate himself to do as he should by contemplating the enticing prospect that once he has admitted Hiker to the castle he will be able to subject Hiker to impermissible servitude and abuse. Misanthrope runs to the door and admits Hiker, whilst making it clear that Hiker may only come in (and avoid certain death) if Hiker agrees to Misanthope's evil conditions. Hiker accepts and enters the castle. 9 Similar points are made by Joseph Carens in his discussion of temporary migrants in Carens (2013, ch. 6). Carens couches his argument in terms of a difference in the responsibility of states within the nation-state system for what goes on within their borders and what goes on beyond them whereas the present argument is centred on democratic agency and responsibility. Carens argues convincingly that if there is no duty on states to promote globally just outcomes then states may not unjustly deprive migrants of rights to secure such outcomes, but if there is such a duty states must pursue it anyway and may not engage in rights deprivation as a means of doing so when they could achieve the goal without doing so.
Notwithstanding Hiker's desperate declarations of consent, Misanthrope has no moral right to subject her to servitude and abuse and, indeed, has a stringent duty not to. It would be perverse in the extreme for Misanthrope to claim that he was morally obliged to offer the regime of subjection and abuse because it was the only way he could get himself to admit Hiker, still less that he was under some duty actually to carry his plan into operation because of the way in which, in his mind, it came packaged with the act of rescue. A different perspective on Misanthrope's actions might be that of a third-party observer. Observer, witnessing the scenario, and sincerely believing that Misanthrope will not be able to get himself to admit Hiker without the incentive of subjection and abuse, may concede that admission on Misanthrope's terms is better for Hiker than dying in the cold. But Observer's disinterested recommendation that this happens should not be taken either as exculpatory of Misanthrope or as a view about what justice requires, all things considered.
In the case where Hiker is at the door, Misanthrope actually has two stringent duties: a duty of rescue and a duty not to violate Hiker's rights. Either (a) those rights are of a kind that cannot be waived by Hiker or (b) the conditions are such that Hiker cannot give valid consent to waiving them. 10 Either will do. Still, we can imagine variants with slightly different features. One dimension of difference concerns the duty of rescue: there might be cases where Hiker is not facing imminent death, but rather a degree of cold and discomfort such that Misanthrope is not under a stringent duty of rescue but, rather, where rescuing is a thing a good person would do, but is not obliged to. 11 We can also imagine cases where the Bprice^exacted on Hiker by Misanthrope -and in anticipation of which he actsis less severe and does not include impermissible levels of subjection and abuse but rather that Hiker agrees to act as an unpaid housekeeper for a period of time or to cook a fixed number of tasty meals for Misanthrope. Variations on the story may therefore parallel some other versions of the openness-rights tradeoff. So, for example, the case where there is a stringent duty of rescue but where the price exacted on Hiker does not involve a violation of Hiker's basic rights, might parallel a case where temporary migrants were denied access to some co-operative welfare schemes. If, conversely, there is no stringent duty to help, but a strict duty to avoid long-term unequal treatment on the territory, we may have something somewhat analogous to the liberal inclusivist view of the Gastarbeiter issue, as found in Walzer. In both cases, however, it can be hard to be precise about the parallels, because so much hangs on a better specification of what the duties, rights and thresholds for treatment actually are.
An obvious question is whether the example scales up to cases where the choosing agent is the democratic electorate of a whole country. There seems to be no difficulty in principle to scaling up the case beyond one in which a single agent chooses. If Misanthrope lives in the castle with his twin brother Scrooge, of identical psychological disposition, and the decision to rescue Hiker is made after discussion and agreement between them, it does not seem hard to conclude both that they have an obligation to rescue and that they cannot subsequently rely on their unjust but motivating desires to justify a regime in which post-rescue Hiker's rights are 10 I assume then that where someone has a stringent duty of aid but where the beneficiary agrees to some condition in order to obtain that aid, perhaps because neither beneficiary or rescuer or both believe that the duty is stringent (even though it is), that condition does not result in beneficiary being subject to a binding duty even though consented to. Some readers of this paper have not shared this intuition. 11 Apparently linguistic intuitions vary about whether it is conceptually possible to rescue someone from a situation not involving serious harms or evils. I think the scope of Brescue^is broader than this and that one can rescue someone even from a boring conversation. The duty of rescue is, obviously, much more constrained than this.
violated. There are more complex cases, though, perhaps more analogous to the reality of electorates with mixed moral views. So, consider one where Philanthrope is the more justiceinclined twin brother of Misanthrope, but where Philanthrope and Misanthrope need to cooperate because the gate is so heavy that a single person alone cannot open it or where it is locked with some electronic device that needs both their passwords. In such cases Philanthrope may (indeed perhaps must) concede Misanthrope's demand to get to exploit Hiker in order that Misanthrope co-operates to get Hiker out of the snow. Philanthrope's pragmatic agreement to this hardly gives moral licence to Misanthrope to exploit Hiker in the future, assuming that Philanthrope had no authority to set aside Hiker's inalienable rights. At best we could say that Philanthrope may under some circumstancesperhaps to ensure Misanthrope's co-operation in repeat Hiker-rescue cases (assume this is a weekly occurrence)have to tolerate Misanthrope's exploitation of Hiker, but this in no way detracts from the wrongness of the exploitation.
Connections between individual and collective responsibility become more diffuse and tenuous as we scale up to state/democratic electoral level, but I see no in-principle objection to the idea that democratic states and their electorates can be responsible for their decisions in a similar way to other corporate bodies with structured decision-making procedures such as companies and clubs. Democratic states can incur debts and be responsible for their discharge and their members can appropriately feel shame and guilt at their acts of wrongdoing and pride in their achievements. Like other agents, individual and collective, they are bound by requirements of justice, including ones to shape the pursuit of their own ends in ways that respect the rights of others. Like other agents, they act in bad faith when they represent their own unwillingness to act in accordance with the demands of justice as an objective constraint which justice itself must bend to accommodate.
There are parallels between the present case and G.A. Cohen's critique of some justifications for Rawls's difference principle (Cohen 2008, ch.1). In that case, the claim that more talented members of society should be paid more because otherwise they would not deploy and develop their talents in ways that benefit the least advantage, loses its moral force when the reason the more talented will not work for less turns out to be their unwillingness rather than their inability to do so. In an actual society, our best option may be to tolerate the unjust greed of the more talented because doing so is practically necessary to secure the best outcome for the least advantaged, but it distorts our moral understanding, specifically, our understanding of justice, to represent a pragmatic compromise with unjust attitudes as exemplifying what justice itself demands. So too would it distort our moral understanding to represent a sacrifice of the rights of migrant workers as required by justice. I return to the relationship between the demands of justice and the politics of the Bsecond best^below.

Does Voluntary Acceptance Make a Difference?
One argument in favour of giving labour migrants inferior status is the possibility of voluntary acceptance. If migrant workers believe they will gain from migration and if they prefer, as consenting adults, to live on the territory of a new state with inferior status to natives to alternative of remaining in their country of origin, what could the objection be? In a recent paper, Kieran Oberman has argued that even a condition of permanent alienage would be morally acceptable if it is true that states have the discretion to exclude would-be immigrants and so can choose to admit on conditional terms (Oberman 2016). Though Oberman rejects the unilateral right of states to exclude because he supports the idea of a human right to free movement and settlement across borders, his arguments have relevance here because those arguing for the openness-rights trade-off support state discretion over immigration.
Although Oberman holds that there is a presumption in favour of equality of treatment for all those settled on the territory of the state, he argues that presumption can be overridden where there is a morally relevant difference among persons. In the present case, such a morally relevant difference between settled foreigners and birthright citizens exists, namely that the different situation of the settled foreigners is a consequence of their voluntary choice. He argues plausibly that Bin a variety of contexts, consent is thought sufficient to prevent injustice.^(2016, p. 101) Although Oberman concedes that there are cases where consent is insufficient to outweigh concerns about injustice, he argues that these are instances where there are great risks of exposure to harm or where it is clear that people have no good reason to give their consent. Neither of these considerations applies in the this case, after all: There are other things in life besides citizenship rights. Migration can allow people to better advance their careers, pursue romantic relationships, be close to friends, or enjoy new surroundings. …. When exclusion is the alternative to permanent alienage, permanent alienage can be in a migrant's best interests. (2016, p. 103) There are some significant differences between what Oberman is saying and Milanovic's view. First, Oberman is explicit about the way in which consent can make a normative difference to the relations among persons whereas Milanovic merely invokes the revealed preference of immigrants in a way that suggests but does not affirm consent. Secondassuming now that the revealed preference argument does impute consent -Oberman explicitly rejects a claim that Milanovic appears committed to concerning the conditions under which consent is valid. Milanovic's argument suggests that the revealed preference even of disadvantaged economic migrants reflects a choice with normative weight whereas Oberman concedes that Bdesperately poor economic migrants have no reasonable alternative^(2016, p. 100) before rather optimistically claiming that governments have the capacity to distinguish sufficiently reliably between voluntary and involuntary migrants. His attitude to a specific case also gives some indication of his thinking: BThe European guest worker programmes of the 1970s (which became permanent alienage programmes since visas were routinely renewed) might also have been beyond the limits to consent given the severity of the harms involved.^(2016, p. 103) Presumably, though, both Oberman and Milanovic would agree with the general point that we can only impute consent to genuinely voluntary migrants, even though they might differ on what we should take to count as voluntary acceptance.
Though Oberman's argument is superficially plausible, I believe it to be flawed in two respects. The first is that it neglects the manner in which states normally regulate contractual relationships with potentially decisive lifetime significance for people. The second is that it assimilates the relationship between a settled resident and the state on whose territory they live to a simple contractual relationship between those parties. Neither the neglect nor the assimilation is sustainable.
Take the neglect issue first. Some relationships involving consent potentially make people extremely dependent and vulnerable. One example of such a relationship is legal marriage, another is employment. Although in the cases of both marriage and employment it is important to give people opportunities to choose freely, it is also important that people within those relationships be protected from abuse, exploitation and domination. To this end, the state may regulate the marriage contract, limit the extent to which voluntary consent to other things, such as sex, can be be imputed from consent to marriage, legislate for divorce, make provision for a fair division of assets in case of divorce, and so on. In some of those cases, if a state fails to regulate by, for example, interpreting consent to marriage as ipso facto consent to sex and thereby making it impossible, legally, for wives to be raped by their husbands, it thereby acts unjustly. Similarly, in the case of employment, many states legislate for a minimum wage, for vacation rights, for rights to fair and decent treatment in the workplace and against unfair dismissal, among other provisions. In the case of neither marriage nor employment do we consider that consent is sufficient to determine all the relevant dimensions of the relationship. Moreover, on the general principle that nobody can be forced to do what they are unfree to do, such state regulation often bolsters the position of vulnerable parties such as workers who might otherwise consent to work in unduly hazardous conditions (Cohen 1988). Uprooting your life and going to live in another jurisdiction is a life-changing decision at least as fateful as getting married or taking a job, and it seems reckless simply to apply general intuitions about consent and status to such a case without noticing that we think that they are not a sufficient guide in these other cases.
The second problem is that is is wrong to assimilate the relationship between a sovereign state and a settled foreign resident to that between two parties consenting to a relationship. In a normal contractual relationship between two parties their agreement settles the terms of their future relationship, at least in the matters covered by the contract. Should one party wish to change the terms of the relationship in the future they must do so by agreement and should disputes arise as to the meaning of the contractual terms in the future, each party normally has access to an impartial legal system or means of arbitration in order to resolve the disagreement. This is because unlike in the normal contractual case, the state party to the state-immigrant relationship can unilaterally change the relationship in fundamental ways after the immigrant party has made a costly and difficult-to-revoke decision to commit. The practical implications of this have been rendered extremely vivid in the case of EU citizens living in the UK (and UK citizens living in the EU) who migrated and settled in another country under a set of assumptions about what their rights to live, work and be with family would be. It turned out those assumptions were unfounded, and they may soon be consigned to a very different legal status that the one they reasonably expected to have. Not only can we not presume that voluntary migrants consented to this reduced status, the fact that sovereign states have the option of adjusting settled migrant residents' status at any time makes it unclear what, specifically, the supposedly implied consent is implied consent to. One possible reply to this point might be to argue that voluntary migrants have consented to a status which includes an understanding that the terms on which they are present may at any time be altered by the state. Because of the vulnerabilities that this gives rise to and the way that it undermines any idea of stable and legitimate expectations, I find this reply unpersuasive.

Which Rights May We Not-Unjustly Restrict?
I have argued that if there are rights to which people are entitled as a matter of justice, such rights may not, consistently with justice or as a demand of justice, be overridden in order to bring about a more just distributive outcome in cases where the obstacle to bringing about a state of affairs where both the distributive outcome and the rights are respected is the unjust attitudes of the agent which is in a position to realise the relevant outcomes. That way lies incoherence. However, this raises the question of, on the one hand, which rights migrant workers are entitled to, as a matter of justice, and, on the other, which rights are within the gift of states to grant or withhold. To some extent, we can leave the matter open and present the issue in conditional form, so that if there are any such entitlements as a matter of justice, then they may not be set aside. Though, logically, such a position is consistent with there being no such rights, I believe that there are some, even when we grant the assumption that states have the discretion to admit or to refuse would-be immigrants. That is to say, if they choose to admit immigrants such as temporary migrant workers, then those people will have some rights that may not be waived and will plausibly acquire more over time.
In specifying the rights to which migrant workers are entitled, Joseph Carens's account seems basically correct (Carens 2013, ch.6). The most obvious set of rights in this category are general human rights, such as the right to integrity of the person and plausibly rights such as the right to emergency medical care. If anyone is present on the territory, they have a claim to have such rights respected and they may not be set aside as a condition of entry. More negotiable are some social rights and labour market rights. Some social rights, such as the right to access unemployment benefits or the right to more extensive healthcare seem like the kind of thing that a state might not unjustly restrict access to. One way in which this might happen is by providing these benefits as part of a contributory scheme, according to which only those who have paid in are entitled to benefits. This may or may not be superior to a noncontributory scheme such as the UK's National Health Service, in efficiency terms, but it is hard to see that it is an unjust mode of provision. Since migrants who have not been in the country very long will not have contributed to the scheme, they will typically be unable to access its benefits. In the case of non-contributory schemes, it also seems not unjust to require that migrants who have been of the territory for a short time pay some fee as a condition of access. What does seem problematic is any requirement that migrants pay into a scheme that then denies them the ability to benefit or have their contributions reimbursed.
Labour market rights, including laws governing health and safety at work pose some distinct issues. Whilst in principle there seems to be no objection, if states have discretion to admit or refuse, in issuing visas that restrict workers to some particular industry or even employer, in practice, given that states have responsibility for protecting the general human rights of those on their territory, and given that such restrictions are know to render migrants more vulnerable and exposed to sundry abuses, caution is needed in practice. I cannot see any justification for exposing migrant workers to lower health and safety standards than citizens are. Versions of the openness-rights trade-off argument such as that advanced by Martin Ruhs, where the trade-off is precisely between social and labour-market rights and openness seem, then defensible in principle, depending on exactly which rights are limited, because openness is promoted without violating any right to which the migrant has a membership-independent claim in justice. 12 The big question, then, concerns membership rights. This is where Milanovic's proposal that migrants should be accorded an inferior set of rights becomes problematic if, for example, something like Joseph Carens's social membership theory of citizenship is true or if theories that argue for membership for all those who are subject to the law and hence the authority of the state and its coercive power are true. I have already rejected the idea that permanent alienage can be made legitimate by taking the fact of voluntary migration to imply consent to an inferior status. If we cannot impute such consent, then migrants who have been resident on the the territory for a sufficient period of time have a claim in justice to citizenship. For them the applicable legal principle (for which there may be several different underlying justifications) will be that Citizenship should be open to everyone who has settled within a state's territory for a certain number of years. 13 If something like this it right, then Milanovic's version of the openness-rights tradeoff is unjustified because it depends upon the citizens of liberal democratic states denying to many migrants something to which they are morally entitled, and something which they could respect whilst also achieving openness, but which they choose not to.
There remains a further, if somewhat disturbing, possibility, which I cannot explore fully here. This is that democratic states could implement the openness-rights trade-off by deliberately not triggering the temporal residence basis for citizenship. In other words, they could admit migrants from poor countries and thereby achieve some of the distributive benefits of greater openness, but issue them with strictly time-limited visas. Such a scheme is advocated by, among others, the economist Dani Rodrik who writes: Rich nations would commit to a temporary work visa scheme that would expand their total labor force by no more than 3 percent. Under the scheme, a mix of skilled and unskilled workers from poor nations would be allowed to fill jobs in the rich countries for a period of up to 5 years. To ensure the workers return home at the end of their contracts, the programs would be supported by a range of carrots and sticks applied by both home and host countries. As the original migrants return home, a new wave of workers from the same countries would replace them. (Rodrik 2011, p. 268) It is hard to say that such a proposal is unjust to any individual, if we make the assumption that states have the right to control immigration. To be sure we can envisage circumstances where even migrants who are present on the territory for less than Rodrik's 5-year period might acquire rights to stayfor example, they might form an intimate relationship with a citizen, they might have a child with a citizenbut these would plausibly be only a minority of cases. If temporary visa schemes do not deny those migrants something they have a right to as a matter of justice, then they are outside the scope of the critique made in this paper. An important objection to them, though, would be that they create a permanent layer of persons on the territory with inferior legal and political status, albeit a layer of variable membership. If the numbers are smalland Rodrik uses a figure of under 3%then this may not be too troubling. Larger numbers, such as those that we find in the Gulf states, do risk creating a society which seems to lack a quality of relational equality that is normally a characteristic of democracy, a quality that is a public good to its members.
6 Is There Really a Big Disagreement Here?
I have argued that it is wrong and unjust to trade-off rights to which migrants are morally entitled in order to secure better global distributive outcomes. But Milanovic sees such trade-offs as a second-best solution, but one that is morally mandated given that the preferences of the citizens of wealthy states are a constraint on policy. Ruhs, though less cosmopolitan in his outlook, than Milanovic seems to agree. As Milanovic puts it It is indeed true that if we lived in a different world where there was a much greater willingness of the populations and governments in the rich countries to accept the idea of free migration of labour, the first-best solution would be precisely to allow such migration and to a treat all residents equally, regardless of their origin. (2016, p. 153).
This apparent concession conceals a subtle but fundamental difference in outlook between those philosophers who see justice as Bthe first virtue of social institutionsâ nd who make a sharp distinction between the moral dimensions of rightness and goodness and others, (Rawls 1971, p.3) including most economists (but some philosophers) who, insofar as they allow for rights, fold them into an all-things-considered consequentialist outlook. For the former group, the denial of rights, though sometimes justifiable, is something to be contemplated only when strictly necessary to avoid catastrophic outcomes. Moreover, since their focus is on our moral agency and our duties both as individuals and as collectives, taking the step to deny rights for better outcomes, when we could achieve those same outcomes by acting otherwise, involves the commission of injustice. For the latter, though it is a truism that a set of consequences in which rights are respected would be better than one in which they are not (hence Bsecond-best solution^) the moral imperative, viewed from the detached perspective of the policy-maker, is to achieve the best outcome on offer given all constraints, including the attitudes of democratic publics.

Conclusion
I have sought to shine some philosophical light on the openness-rights trade-off proposal made by a number of economists. To reiterate, all versions of the trade-off proposal assume that states have the right to regulate immigration. Though that assumption may be unfounded it is widely believed by politicians, the general public of wealthy democratic societies and by the economists who have explored and advocated the trade-off. I have argued that where the proposal involves rights that immigrants are entitled to as a matter of justice then it fails as a recommendation for what justice requires, because it is perfectly possible for democratic electorates to choose both to admit more migrants and to respect the rights of migrants to things like political membership. That they choose not to, because of their unjust attitudes, is not the kind of objective constraint that should ground a genuine trade-off. Where the trade-off involves rights which cannot be waived, then the proposal looks attractive from the point of view of a detached observer, but is incoherent as the action of a democratic public seen as a moral agent. The mere fact that migrants choose to migrate cannot be taken as evidence of their consent to an inferior legal status. However, trade-offs of less important rights are sometimes permissible and temporary visa schemes may not be unjust as such.