The Open Borders Claim in a Nonideal World

Although many writers in political philosophy argue for open borders, the question of what this entails for nonideal worlds has not yet been adequately addressed. Two diverging positions can be discerned in the literature. On the one hand, writers such as Joseph Carens emphasize that borders should only be open under ideal circumstances, and that the Open Borders Claim does not at the present moment justify pursuing a political program. On the other hand, writers such as Andreas Cassee and Kieran Oberman argue that states should open their borders immediately to a great extent. This paper argues against both positions, and sets out an intermediate approach.

The Open Borders Claim seems at first sight to be a useful tool to evaluate current migration regimes, or to justify political programs or civil protests. If the claim is valid, then the extensive restrictions that liberal states impose on migration seem to be highly problematical. The Open Borders Claim, for example, brings to mind those "Freedom of Movement is Everybody's Right" banners in the hands of demonstrators arguing for the rights of asylum seekers. In fact, the No Border Network, an actor from the political left, aims at undermining the reality of closed borders. 2 Thus, the philosophical question of whether the Open Borders Claim is justified seems highly relevant when it comes to answering such questions as whether we should promote political attitudes like these, which party we should vote for in national elections, and whether we should give financial support to certain NGOs.
However, given how our world looks today, is there really a direct path from the Open Borders Claim to endorsing a certain political program? Academic proponents of open borders have not yet sufficienty approached this question, but the positions that can be glimpsed from their mostly vague statements show that they disagree amongst themselves. Some believe that, even if border restrictions are imposed within the framework of a world like ours (a world with huge and largely unjust inequalities between individuals and states), the claim justifies the normative statement that states should throw open their borders to a great extent. Others reject such a view, claiming instead that open borders form a future utopia, taking place in an ideal world without huge economic differences between states; for them, it does not justify a political program within our world.
In what follows, we discuss both positions in order to gain an insight into what the Open Borders Claim implies for nonideal worlds. In section 1, we provide the conceptual basis for our discussion by taking a closer look at how Joseph Carens approached this issue in his 2013 book The Ethics of Immigration (EoI). The different levels of theorizing we find in Carens help us distinguish a weak version of the claim, held by Carens, from a strong version, held in particular by Andreas Cassee and Kieran Oberman. Both versions are open to serious objections. In section 2, we present two arguments for rejecting the strong version; in section 3, we discuss two issues that give reason to doubt the weak version. We conclude in section 4 that an intermediate approach is necessary, and we offer some ideas that could be a first step towards formulating a nonideal theory of open borders.

Two Versions of the Open Borders Claim
In EoI, Carens gives a two-fold argument for open borders (see ch. 11). First, he argues that closed borders now have an effect comparable to the effect that feudalism had some centuries ago: some people are born in rich states and are therefore privileged by birth when compared to those who are born in poor states. Thus, equality of opportunity requires opening borders for every human being.
Second, Carens claims that every justification for Art. 13, § 1 of the Universal Declaration of Human Rights ("Everyone has the right to freedom of movement and residence within the borders of each state.") also justifies the human right to freedom of movement around the whole globe. Hence, anybody who affirms Art. 13, § 1 also has to affirm the idea that freedom of movement across borders is a human right, too.
However, Carens also qualifies the Open Borders Claim in the further course of his argument. While arguing that open borders are just in principle, he does not intend his claim to have any significant political implications. Rather, he restricts the scope of his argument thus: "Fundamentally, however, the open borders argument itself is not intended to advance a specific program of action" (EoI, 296).
Instead, Carens emphasizes that the Open Borders Claim is designed for ideal circumstances only: "What would a just world look like? Iʼm not sure of all the particulars, but I am reasonably confident of this. If a just world had states, they would be states with open borders" (EoI, 287).
As such a utopian ideal, open borders constitute more a general political aim of the future than a right that can be claimed in the present. While, according to Carens, there is a duty to aspire to that ideal in the course of history, there seems to be no duty to enact open borders at present: While I do think that a just world will be one in which people are largely free to live where they choose and in which there is relative economic equality among places and people, I am not certain that the best way to move in the direction of that world is to increase immigration to rich democratic states as much as possible (EoI, 296).
In making these claims, Carens seems to be following John Rawls's distinction between two types of theory, ideal and nonideal theory (Rawls 1971). In ideal theory, we imagine an ideal world that solely consists of states that act according to principles of justice and that are equipped with sufficient resources. Thus, in ideal theory, we can assume that there is only a modest degree of economic inequality between states: most sources of inequality do not exist. In nonideal worlds, in contrast, not all states act in accordance with what justice requires them to do, and not all states have sufficient resources. Thus, there is huge economic inequality between states, comparable to the situation in the world today.
It is important to emphasize that elaborating a nonideal theory is not the same as limiting theory to what is feasible or what can be realistically achieved. If we claim that certain states should act in a certain way in nonideal circumstances, then we can concede that these states will probably act otherwise, due to the impact of self-interest or populism. Theorizing on a nonideal level is therefore to ask what justice requires a certain agent to do, given that some other agents do not comply with the demands of justice, while disregarding how likely it is that the agent in question will in fact act upon that requirement.
Although it is not entirely clear whether Carens applies the Rawlsian distinction between ideal and nonideal theory consistently, 3 it seems quite appropriate to read him this way: in an ideal world, states must open their borders (and can only refuse to do so in very exceptional circumstances; see EoI, 293).
In contrast, there is no demand on states to open borders in nonideal worlds, and the Open Borders Claim does not justify any political program given a world like ours.
In order to gain a more comprehensive view of Carens's approach, we should bear in mind that Carens addresses different levels of reasoning in a second way. As we understand it, Carens's approach is characterized by two main distinctions, the first between ideal and nonideal theory, and the second (and no less important) between two ethical assumptions, the so-called Conventional View and the Open Borders Claim. Most of EoI does not make use of the Open Borders Claim that we have presented here. Instead, Carens explicitly follows the Conventional View, which is the view that states have the discretionary right to control their borders, with exceptions for cases like family reunification and asylum. Carens assumes that most people today share the Conventional View, and that it would therefore be a mistake to base arguments unnecessarily on premises that contradict the Conventional View. Only the last chapters of EoI follow the path of rejecting the Conventional View. Thus, one part of Carens's reasoning is based on the Conventional View, and another is dedicated to proving that the Open Borders Claim is justified.
These two distinctions -between ideal and nonideal theory, and between the Conventional View and the Open Borders Claim -make it necessary to keep apart four subtypes of theory that arise from combining the two theory types and the two ethical assumptions: Carens devotes most of his book to TSnonideal,CV: he investigates what common democratic principles imply, given the world how it is now, with the Conventional View being among these widely shared democratic principles (chs. 2-10). There is no reason to believe that TSideal,CV differs substantially from TSnonideal,CV. Thus, we can assume that, according to Carens, the same principles apply in TSideal,CV as in TSnonideal,CV, perhaps with the difference that some of these will lose their importance when transferred to TSideal,CV. As we have already indicated, the chapters concerning the Open Borders Claim are intended to spell out the principles of TSideal,OBC (chs. 11-13), and exclusively so. In the conclusion of his book (ch. 13), Carens discusses how his claims from the previous chapters about the rights of immigrants also apply (with some minor qualifications) in an ideal world with open borders. Thus, the "Conclusion" definitely addresses TSideal,OBC, too.
However (and this is of great significance for our purposes), Carens largely disregards TSnonideal,OBC. What we know is that there is a duty in TSnonideal,OBC to overcome the unjust framework conditions by bringing about an ideal world -but it is difficult to say which acts promote that aim. The crucial question, however, is: given that we do not live in an ideal world and that we do not have the means to bring about an ideal world immediately, and given also that the Open Borders Claim is valid, which migration policies should we endorse?
One objection that could be made here is that it is not necessary to know about TSnonideal,OBC. Most people believe that the Conventional View is true, and, since Carens's claims regarding TSnonideal,CV are already demanding, it could, given the restrictive migration policies of practically all states, be superfluous to address TSnonideal,OBC. We are convinced that this objection is unfounded, however.
Although providing arguments based on premises that we ourselves think to be false might be rational (and even morally required out of respect for other people's convictions), we cannot ignore the question of what the convictions that we ourselves hold to be true imply. If we have built a theory on assumptions that we hold to be false, then we should abandon that theory; there is no sense in affirming a theory whose premises are false. Thus, if we believe that the Open Borders Claim is justified, and if we want to know which policies we should endorse, then we have to ask what the Open Borders Claim implies for nonideal worlds, regardless of the fact that other people hold the Conventional View.
Indeed, there is some evidence that Carens does actually have an approach to TSnonideal,OBC, although the approach is drowned out by the main line of his arguments. In short, Carens seems to believe that TSnonideal,OBC has more or less the same content as TSnonideal,CV. Let us mention two observations that support this interpretation. First, Carens discusses TSnonideal,OBC very briefly in the conclusion of EoI, when he considers certain policy fields. Under the heading "What is to be done?", he writes about irregular migrants and states that, given that the Open Borders Claim is justified, we should use the same principles as he has developed for the Conventional View: Those who accept the open borders argument should support the same policies. [...] So, the analysis in that chapter also offers relatively concrete guidance about how to act in the world, guidance that should be helpful not only for those who accept the state's right to control immigration but also for those who believe in a more demanding vision of justice (i.e. open borders) but want to promote justice as much as possible within the limits of what is feasible (EoI, 295).
His view, then, seems to be this: given the world as it is now, everything that is argued for within the Conventional View is also valid if the Open Borders Claim is justified. The Open Borders Claim just adds further moral duties, especially the duty to bring about a world where open borders are possible, as well as perhaps some other duties. Miller (2015) has called "das Carensproblem": like the problem of how to reconcile the two books of Adam Smith (widely known as the Adam Smith problem), there is, according to Miller, the problem of how to reconcile the part of Carens's approach that challenges the Conventional View with all those chapters that assume that the Conventional View is valid. The Carensproblem can be restated as a set of three mutually exclusive statements:

Second, let us introduce what David
(S1) The arguments in chapters 2-10 of EoI presuppose the assumption that the Conventional View is valid.
(S2) The arguments in chapters 2-10 of EoI are sound.
(S3) Chapters 11-13 of EoI prove that the Conventional View is not valid.
It seems that Carens can reasonably only hold two of the three statements to be true. If S1 and S2 are true, Carens has to concede that the Conventional View is valid. If S1 and S3 are true, Carens has to concede that his arguments in chapters 2-10 only have the strategic purpose of pursuing some opponents, but he himself dismisses the idea that these arguments are sound. If S2 and S3 are true, however, it cannot be true that chapters 2-10 presuppose the Conventional View, if we take "an argument presupposes x" in the sense of "if x is unjustified, then an argument that presupposes x fails".
Interestingly, in his reply to Miller, Carens does not question one of the statements in general, but nevertheless rejects the view that the Carensproblem raises a serious objection to his position (see AAPT). In particular, he does not argue that the validity of the Conventional View is irrelevant for his arguments in general (which would reject S1), and nor does he concede that the claims from chapters 2-10 have a merely strategic value (which would reject S2). Instead, he seems to pursue a further strategy, namely modifying S1 and S2 in a way that is compatible with S3. Most of his defense is concerned with the fact that the Open Borders Claim is not intended to be a principle that guides action in a nonideal world, although many of his claims in the first part of his book are. That strategy could be restated as: (S1') Chapters 2-10 of EoI are largely independent of the question of whether the Conventional View is valid as long as we work within a nonideal framework; and presuppose the Conventional View, as soon as we work within an ideal framework.
(S2') Most of the arguments from chapters 2-10 of EoI are sound when faced with a nonideal world.
(S3) Chapters 11-13 of EoI argue that the Conventional View is not valid.
In consequence, Carens seems to hold the view that there is only a minor difference within nonideal theory between those who presuppose the Conventional View and those who presuppose the Open Borders Claim. This, however, implies that Carens actually has a position concerning TSnonideal,OBC: for Carens, TSnonideal,OBC contains more or less the same principles as TSnonideal,CV. Of course, the extent to which Carens admits deviations between TSnonideal,OBC and TSnonideal,CV remains unclear. 4 However, there is for him at least broad accordance. Let us call that view the "Weak Version of the Open Borders Claim" ("OBCweak").
(OBCweak) In nonideal worlds, the Open Borders Claim does not have substantial implications for how to act right now that are not also implied by the Conventional View.
Why does Carens hold OBCweak? He does not offer any elaborated argument to support such a view. As far as we can see, Carens briefly mentions in the course of his reasoning two prima facie arguments in support of OBCweak. First, most of the principles that Carens defends in EoI are concerned not with the admission of immigrants, but with the rights of those who have already migrated. Since the Open Borders Claim addresses the question of admission, it is plausible to assume that the rights of immigrants are not affected by the Open Borders Claim (ch. 13). Second, the Open Borders Claim implies not an absolute right to migrate, but rather a right that can be restricted under certain circumstances. Since "[p]olitical communities require relatively stable, intergenerational populations in order to function effectively over time" (EoI, 287), it might be justified for states to close their borders when there is a threat of mass migration. As long as there are great inequalities between states, mass migration is indeed a plausible result of open borders. Thus, it seems that states may close their borders, just as they may do so within the Conventional View.
In the philosophical debate, some defenders of open borders seem to pursue a more ambitious aim than Carens. They hold that the Open Borders Claim has clear implications for nonideal circumstances, although there might be reasons not to open borders completely. Kieran Oberman and Andreas Cassee claim that, as a human right, the right to open borders should be in force for all human beings at all times and in all places, and can be restricted for exceptional moral reasons only. From their point of view, Carens's modest conclusion is striking, since it denotes global freedom of movement as a human right, but diverges greatly from what that would imply: the language of human rights is usually used precisely when we want to make demands that are to be fulfilled immediately, independently of contingent circumstances and the full compliance of other actors. It is fair to say that the idea of human rights as minimal standards of justice is even developed specifically to address normative problems in a nonideal world. However, although Carens holds that there is a human right to freedom of movement across borders, he apparently does not hold that individuals can claim that right under the present conditions. 5 Thus, Oberman claims that restrictions on immigration "might be justified in extreme circumstances" only (Oberman 2016, 33; see also 46), and that limiting the right to migrate should follow the same rules as for other human rights. He explicitly criticizes Carens for emphasizing that open borders are a mere ideal, and for not using the language of human rights consistently (Oberman 2016, 33 [fn. 11]).
A similar view is held by Cassee, who argues that just principles to regulate immigration are those principles that all humans would rationally choose in an original position (in Rawls's sense) -that is, a position from which people lack knowledge about who they will be (whether they will be born in a rich 5 A deeper issue seems to be at stake here: Carens does not reject the common assumption that freedom of movement within a state is a human right in the sense that it is possible to appeal to that right in order to justify concrete claims under nonideal circumstances. He then holds that there is a strong analogy between freedom of movement within the state and global freedom of movement. This analogy seems to justify the conclusion that global freedom of movement is also a human right in the sense that it is possible to appeal to that right in order to justify concrete claims under nonideal circumstances. However, Carens does not draw that conclusion, and he fails to explain from where the difference comes between the two rights of freedom. Such an explanation is suggested in Hoesch (2017a), who defends the view that global freedom of movement is a human right in the sense that it is a moral prima facie right that every human being has only because she is a human being; but only national freedom of movement is a human right in the sense that it is a right that deserves strong priority when it comes to conflicts between moral considerations. See also Düring/Luft, chap. 10 in this volume. or in a poor state, what talents and preferences they will have, etc.). Cassee claims that people in such a position would choose principles that are very close to completely open borders. With regard to applying these principles, he writes: It is basically the right of every human being to move freely on the surface of the planet that we inhabit together, and any exception to this principle requires justification by other equally weighty claims (Cassee 2016, 279 [our translation]).
Cassee then implicitly concludes that a state is obliged to open its borders to a great extent in order to increase overall justice: Small steps towards a more liberal migration policy are not [...] only small reductions of a great injustice, but potentially also pave the way for the complete overcoming of unjustified coercion, which is exercised daily against potential immigrants (Cassee 2016, 281f. [our translation]).
As can be seen here, Cassee views restrictions on immigration even in a nonideal world as unjustified coercion, and every step taken towards open borders as a reduction of injustice. That means, in turn, that justice requires the immediate opening of borders, even in nonideal worlds. Thus, Cassee views concrete activities in the actual world in light of the Open Borders Claim; in contrast to Carens, he insists on advancing a program of political action.
As we see it, Cassee and Oberman both tend to fill the gap of TSnonideal,OBC with principles from TSideal,OBC: for them, states have to open their borders immediately and to a great extent, with every single exception needing substantial justification. We will call that the "Strong Version of the Open Borders Claim" ("OBVstrong"). The Strong Version is perhaps best characterized by two claims: that every human being has the right to enter a state as long as that state does not offer specific reasons for excluding her; and that, given the current restrictions on immigration, every step towards liberalizing migration policies is a step towards reducing injustice.
(OBCstrong) In nonideal worlds, the Open Borders Claim implies that states should open their borders to a great extent, with every single exception needing substantial justification.
Surprisingly, and contradicting what we have said above, Carens also seems to endorse OBCstrong in one passage of EoI, when he claims that, if the Open Borders Claim were true, he would have to revise his arguments on irregular migrants. For him, the Open Border View would mean that such migrants simply have enacted a moral right that should also be acknowledged in positive law, and that they should therefore not be subject to any constraints. Carens seems to be claiming here that, even in nonideal circumstances, these migrants had the right to migrate, and the Open Borders Claim therefore does justify a certain political program (albeit a program that is unrealistic to be implemented). 6 6 We will return to that claim later.
The distinction between OBCweak and OBCstrong is not absolute. Rather, there is a range of possible versions between the two, these versions being weaker or stronger depending on whether they endorse principles that are closer to TSnonideal,CV or to TSideal,OBC. Indeed, in our view, neither extreme position, OBCweak or OBCstrong, is convincing, so that the difficult task is to spell out an intermediate position concerning the Open Borders Claim in nonideal worlds.
In the next sections, we first propose two arguments as to why OBCstrong fails -hence, these arguments are in line with Carens's approach of not making use of the Open Borders Claim to justify liberal migration policies in nonideal worlds. Then, we present two cases in which in our view the Open Borders Claim does make a difference for nonideal worlds, so that OBCweak also fails. In the conclusion, we outline what we hope is a plausible intermediate position.

Why the Strong Version of the Open Borders Claim Fails
According to OBCstrong, states should, in order to increase justice, open their borders to a great extent, even in nonideal circumstances. The two arguments that suggest that OBCstrong fails, we discuss in the following, begin from the fact that mass migration would be a serious problem. As Carens rightly points out, states need a relatively stable population in order to function (EoI, 287). The threat of mass migration might be a reason to close borders even if we acknowledge global freedom of movement as a human right. The arguments take different paths to show why the threat of mass migration constitutes a general objection to OBCstrong.  Cassee 2016, 253). He is using the method that Rawls outlined in A Theory of Justice here of theorizing on the basis of an original position; in our view, though, it is precisely this method that leads him into a serious problem. In A Theory of Justice, Rawls (1971) wrote about a society where the principles chosen are enforced by law -that is why the parties in Rawls's theory quite rightly presume full compliance. However, in the case of open borders in a world of sovereign states, there is no law that forces the majority of states to comply; hence, the presumption of full compliance is implausible and leads to fallacies.
Note first that for a state to open its borders while all other states keep theirs closed may be regarded as a moral overload. Imagine that state A opens its borders, but states B, C, D … keep their borders closed. What would happen? It is very likely that some of the migrants who have tried to reach B, C, D ... will migrate to A instead. So, the problem is that, even if A remains willing to accommodate all those who want to migrate, it may be impossible for A to keep the state functioning while also accepting the number of people who want to migrate to A. As we have said, mass migration is a reasonable threat and, if we know that only a few states comply with the Open Borders Claim, the risk of mass migration increases dramatically.
Needless to say, states do not have to wait until immigration actually causes serious problems. Instead, states do well to restrict immigration preemptively -they do not have a discretionary right to do so, but may restrict immigration to a level that they believe to be compatible with the long-term functioning of their institutions. Therefore, states should implement immigration laws that follow reasonable principles of selection to allow for a limited number of entries. Instead of opening borders to a great extent, states would be required to allow migration via a selective process of admission.
And, instead of having to justify closed borders in every single situation, states would be allowed to close their borders preemptively. 7 The fact that most states will not comply with the Open Borders Claim gives rise to a further problem.
While opening borders would end the supposed injustice that migrants cannot enter, it would also lead to another injustice: namely, injustice between citizens of different states. Emigrants from B, C, D ... now have the chance to migrate to A, but citizens of A cannot migrate to B, C, D ... What seems at first glance to be an increase in justice (inhabitants of B, C, D are no longer constrained from migrating to A) turns out to be an increase in injustice in another respect (inhabitants of A are disadvantaged

compared with inhabitants of B, C, D). Thus, the fact that B does not open its borders to citizens of A is a reason for A not to open its borders to citizens of B. 8
However, there is an important objection to that argument. It could be argued that, if global freedom of movement is a human right, the fact that B, C, and D do not comply does not justify A's noncompliance. Take the example of the right to religious freedom. Suppose that Saudi Arabia does not provide freedom of religion for German emigrants. Is that a reason for Germany not to provide freedom of religion for Saudi emigrants who live in Germany? Surely not. We should not punish individuals for the moral mistakes that their home state or any third state has committed; and, even if we were allowed to do so, that punishment should not violate human rights.
We believe that, when applied to the supposed human right to global freedom of movement, this reasoning is built on a false understanding of the sense in which global freedom of movement can be seen as a human right. Global freedom of movement is certainly not one of those rights that we by all means need to be able to lead a good life. Even if it is a human right in principle, it is far from being one of the fundamental rights that can never be restricted. Furthermore, freedom of movement differs from other human rights in that, since places are limited, granting the right to some people means withholding it from others (in contrast, we can easily grant freedom of religion to everybody). Given that, it seems plausible to think that states are required to open their borders only to those citizens whose home states also open borders. In summary, the argument shows that justice does not require states to open their borders to a great extent; rather, they should do so only according to general criteria of selection, one of which is reciprocity between states.
Our second argument concentrates on the fact that the Open Borders Claim, if justified via the idea of a human right to cross borders, is valid for everyone, independent of the reasons that she has to migrate. Cassee emphasizes that it is up to every individual to decide whether she has a good reason to settle in another place. As a consequence, it should not be up to states to set rules about who may migrate and who may not. On the other hand, Cassee concedes that priority should be given to those whose safety is threatened (see Cassee 2016, 215); he even emphasizes that the Open Borders Claim is attractive just because it gives weight to the interests of refugees (see Cassee 2017, 40). In the same way, we should acknowledge that family reunification generates a priority claim when it comes to immigration.
However, as Rainer Bauböck has repeatedly pointed out, there is no fixed number of free places for immigrants that can be filled either with refugees or with economic migrants; he diagnoses here a common "fallacy of fixed integration capacities" (Bauböck 2018, 145), and argues that economic migrants might even enhance a state's ability to take in asylum seekers and refugees, e.g. by creating jobs. This makes things even more complex, but it does not change the core fact that places are limited, both for refugees and for other migrants. Thus, the following argument appears to be sound: (P4) If states have limited capacities of admission, are faced with some potential immigrants that have priority claims of admission, and can expect to be faced with further potential immigrants who have priority claims in the future, then states should admit those potential immigrants and so many of them that in the long run they will admit as many as possible of those with priority claims.
(P5) If states should admit those potential immigrants and so many of them that in the long run they will admit as many as possible of those with priority claims, states should in the short run select as many as possible of those who have priority claims and some of those who can be expected to promote its future capacity of admission.
(C) So, all states that are willing to open their borders should, in the short run, select as many as possible of those potential immigrants who have priority claims and some of those who can be expected to promote the state´s future capacity of admission.
As we have already indicated, (P1) is accepted by all participants in the debate. (P2) is held explicitly by Cassee, and (P3) seems to be a highly plausible addition to (P2). (P4) and (P5) are plausible specifications of Cassee's claim to give priority to those immigrants who fear for life and limb. Thus, Cassee should accept the argument. However, borders that are highly selective, as expressed in (C), are de facto relatively closed borders. That means that, under nonideal circumstances, the Open Borders Claim not only justifies but also demands closed borders and highly selective, non-liberal border policies, possibly even including strict deportation policies. 9 The argument proves that the claim that Cassee and Oberman seem to be making -namely, that every action towards liberal migration policies reduces injustice -is not justified in general. Rather, only those liberalizations of migration restrictions that do not worsen the situation of refugees are just. Hence, states should open their borders according to general criteria of admission that select or serve the neediest amongst the potential immigrants.
If that is true, the Open Borders Claim in nonideal worlds demands border policies that are very similar to those demanded by the Conventional View. The reason is that, if the Conventional View is valid, states have an obligation to admit refugees and to take action to extend their capacity to do so. The Open Borders Claim would not substantially alter that obligation. Thus, we are convinced that OBCstrong fails. States should not open their borders to a great extent, but implement general laws of admission that consider amongst other criteria the priority claims of refugees and reciprocity between states.
We have mentioned above a passage where Carens holds OBCstrong with regard to irregular migrants.
Let us briefly investigate what our arguments imply for this case. Carens claims that, even though the Open Borders Claim does not justify policies in nonideal worlds in general, it does so in the case of irregular migrants: if, the Open Borders Claim is valid, they must, according to Carens, not be subject to any constraints on account of their status as immigrants. However, given that states should implement general rules that give preference to the neediest and that recognize reciprocity between states, then why should states give priority to those who were able to cross their borders illegally when it comes to deciding whom to admit? To be sure, irregular migrants acquire the right to stay after some period of time. But this is also true within the Conventional View. As long as irregular migrants have entered the country only recently, any state who is willing to comply with duties of justice would be free to deport some of the irregular migrants in order to admit others who would like to enter and who meet the selection criteria better. Thus, we believe that Carens's general position that the Open Borders Claim does not justify an individual right to enter in nonideal circumstances is also valid in the case of irregular migrants.

Does the Open Borders Claim Make Any Difference in Nonideal Worlds?
According to OBCweak, the principles that apply in nonideal theory are more or less the same principles that would also apply if the Conventional View were valid. Remember that the Conventional View claims that states have the discretionary right to restrict immigration, except when it comes to the reuniting of family members or the admission of refugees. So, the question is whether the Open Borders Claim makes any difference in nonideal worlds or not. Without elaborating too much on the issue, we present two cases that give reason to reject OBCweak.
The first case in which the Open Borders Claim makes a difference is the issue of ordinary admission.
As we indicated in the second section, states are obliged in nonideal theory -in TSnonideal,CV as well as in TSnonideal,OBC -to admit as many persons with priority claims as possible. Nonetheless, there are also some places that can be offered to people who do not have priority claims. This also holds true for both ethical assumptions, the Conventional View and the Open Borders Claim. However, the Conventional View and the Open Borders Claim disagree in how those places are to be allocated among potential immigrants.
If we assume that the Conventional View is justified, then states are morally free to enter into agreements with other states that serve their interests. Think about agreements on free movement between states, such as free movement within the European Union (EU). Given that member states of the EU believe that free movement serves their interests, they are morally allowed to establish rules that guarantee free movement to their citizens. As long as migration between the states that have established free movement does not reduce the possibility to admit refugees, those states are perfectly free to make or not to make such agreements.
Within the Conventional View, free movement can even be used as a tool in political negotiations: states are free to offer free movement to the citizens of any other state with the aim of gaining some advantage in return. The whole history of the EU is full of such negotiations regarding whom should be admitted as a new member of the EU, or whom should be admitted into the Schengen treaty, with interests always being weighed according to the logic of "give-and-take". Carens rightly mentions that there are also normative requirements in the Conventional View that the criteria of exclusion and selection have to meet -states have to reach agreements on free movement that guarantee that nobody will be excluded on account of her race, for example (see EoI,ch. 9). But, as long as those requirements are not violated, states are free to treat citizens from some countries differently to citizens from other countries.
If we abandon the Conventional View, things look different. According to the Open Borders Claim, every human being has a prima facie right to migrate. As a consequence, states are no longer free to offer entry according to whatever preferences they have. Instead, they have to offer entry according to a system of giving priority to some on the basis of objective criteria -this seems to be the only morally acceptable way to respond to the fact that everybody has the human right to migrate in principle, but not everybody can be admitted because of the threat of mass migration. Every criterion of selection that would include discretionary preferences of states would violate the equality condition inherent in all human rights. Note that these priority criteria are not limited to giving priority to uniting family members and to admitting refugees. Every decision on admission has to follow some objective criteria.
Take again the example of the EU. Suppose that the member states of the EU are able to absorb a certain number of immigrants n per year, and n is exhausted by immigrants from other member states of the EU and by the refugees whom they admit. Consequently, those member states deny entry to citizens of third states. It seems that hereby, the member states are violating the rights of those citizens of third states: since there is no objective process of choosing those who are admitted, citizens of third states cannot from the outset make use of their human right to free movement. As a result, given that the Open Borders Claim is justified, agreements on free movement could turn out to be illegitimate, because they hinder an objective selection process that considers all human beings equally.
Of course, as we outlined in section II, one of the objective criteria that states should use is the criterion of reciprocity between states: states should open their borders primarily to citizens of states that are themselves prepared to open their borders. Thus, in some sense, it is legitimate to offer open borders as a return for open borders. However, that does not mean that states are totally free to choose other states in order to conclude agreements of reciprocal free movement. Instead, states have to offer selective migration agreements to all states that are also willing to open their borders reciprocally. It is plausible to assume that there are many states in the world that are willing to conclude agreements of free movement with member states of the EU. That these member states decided to allow free movement among themselves but not with other states is a discretionary decision based purely on self-interest, and is a decision that lacks legitimacy if the Open Borders Claim is valid. Since free movement is a human right according to that claim, it is surely impermissible to negotiate free movement in the context of self-serving agreements -by, for example, offering open borders in return for economic gains.
To sum up, the Open Borders Claim makes a great difference to real-life border politics. It demands that a migration policy based on the interests of states be replaced by one that is based on the rights of potential immigrants; and it suggests that many of the agreements between states could be illegitimate.
Let us turn to the second case that can be made to argue for a difference between the Conventional View and the Open Borders Claim in a nonideal world -a case that is located in the area of the rights and duties of immigrants rather than in the area of admission: namely, integration requirements. One argument in favor of OBCweak is that, as far as the rights of those who have already migrated to a country are concerned, the question of whether they have the right to migrate is not relevant (see section I).
However, is that really true? Think about so-called mandatory civic integration policies -an issue that Carens does not discuss, 10 but one that has gained much attention in the social sciences recently. In the EU, most states today require immigrants to learn the official language and to join classes to learn about the culture, the moral values and the history of the country, with these requirements often being reinforced with penalties. 11 Even within the Conventional View, the obligation to join such classes appears unjustified in many cases (and most EU states violate the moral rights of some immigrants, since they force them to join these classes and punish those who refuse to do so autonomous agents when it comes to those immigrants who are neither refugees nor have a family link that constitutes the right to migrate. In more detail, they could argue: (P1) States have the discretionary right to exclude all potential immigrants who do not have a moral right to migrate on account of being reunited with their family or their status as refugees.
(P2) If a has a discretionary right towards b to do x, a is allowed to conclude agreements not to do x in return for other services performed by b or any third party, as long as the commitment to these services meets a certain set of minimal moral standards s.
(C1) States are allowed to conclude contracts with potential immigrants that do not have a moral right to migrate on account of being reunited with their family or their status as refugees in order to grant admission in return for services that the immigrants will perform, if the commitment to these services meets s.
(P3) The commitment of immigrants to participate in integration classes meets s if the classes observe neutrality, non-discrimination and proportionality.
(C2) Thus, states are free to require immigrants who do not have a moral claim to migrate on account of being reunited with their family or their status as refugees to sign an integration contract, thereby committing themselves to participating in integration classes that observe neutrality, non-discrimination and proportionality.
In consequence, within the Conventional View, integration requirements can be justified for some people, although we should add that the content of the integration contract must be consistent with basic rights and be non-discriminatory (e.g., there must be no requirement to accept a certain religious denomination), with the normative force of the contract diminishing over time and disappearing entirely once the immigrant has acquired the moral right to stay. 13 But things look different as soon as we abandon the Conventional View. Now, there is no discretionary right to permit entry. Instead, states are justified in rejecting entry according to objective criteria only.
But, as soon as they admit entry to someone, it is not the case that states transfer a right that they are free to transfer or not. Instead, in admitting somebody, states are just responding to the existing moral right to migrate. So, states are not free to allow entry only under conditions that they determine at their will; hence, they must not demand immigrants to sign an integration contract that exceeds the requirement that immigrants respect the law. Thus, accepting the Open Borders Claim means seeing integration requirements as unjust in general.

Conclusion
We have distinguished two versions of the Open Borders Claim. The Strong Version (OBCstrong), as defended by Cassee and Oberman, understands the right to cross borders as a valid human right. This right gives support in the actual world for opening borders to a great extent, so that more or less the same principles apply in a nonideal world as in an ideal world. We believe that OBCstrong, understood in such a general way, is unjustified. The danger of mass migration makes it necessary to implement very selective immigration arrangements preemptively.
The Weak Version (OBCweak), which Carens seems to hold, considers freedom of movement across borders as a valid human right in ideal worlds, but does not justify policies in nonideal worlds such as ours. Instead, Carens recommends following more or less the same principles in nonideal worlds that are required if we do not hold the Open Borders Claim. To counter that idea, we showed that the Open Borders Claim does make some difference for theorizing migration in nonideal worlds. As soon as we abandon the view that states have a discretionary right to close borders, states are no longer free to withhold entry from whomever they wish. Instead, every rule of admission must refer to objective criteria that make clear why certain persons are admitted, and others are not. And, when it comes to admissions, states are obliged to treat immigrants like persons with a moral right to migrate.
We are far from having an adequate account of all the details of what we have called theory subtype TSnonideal,OBC, which is the theory about what the Open Borders Claim implies for nonideal worlds.
However, it seems that states should implement general rules of admission that aim at the following four requirements, listed in order of importance: 1. They try to admit in the long run as many of those potential immigrants as possible who have priority claims to migrate.
2. They select amongst those who do not have priority claims according to objective criteria of who should have preference.
3. They aim at reciprocity between states. 4. They seek to offer all human beings as much freedom to migrate as is compatible with the first three requirements.
The question of what the Open Borders Claim entails for nonideal worlds has not yet been adequately addressed. If our reasoning is correct, then the Open Borders Claim demands the cessation of some policies that have up until now been widely accepted, such as negotiating agreements on free movement in return for economic benefits, or establishing integration requirements based on the idea of an immigration contract. However, we did not elaborate on the question of whether the Open Borders Claim itself is justified. Our conclusions will surely motivate to address that question with due diligence once more.