Poverty and Immigration Policy

What are the ethical implications of global poverty for immigration policy? This article finds substantial evidence that migration is effective at reducing poverty. There is every indication that the adoption of a fairly open immigration policy by rich countries, coupled with selective use of immigration restrictions in cases of deleterious brain drain, could be of significant assistance to people living in poor countries. Empirically there is nothing wrong with using immigration policy to address poverty. The reason we have to reject such an approach is not empirical but normative. People have human rights to stay in their home country and to migrate elsewhere. Counter poverty measures that require people to move or to stay are likely to violate these rights. Everyone should be free to migrate but no one should be forced to migrate. Using immigration policy to address global poverty, in place of alternatives, fails on both these counts.

W hat are the ethical implications of global poverty for immigration policy? In a world in which 1.2 billion people live at the margins of survival, on less than $1.25 (PPP) a day, poverty is arguably humanity's most pressing problem (United Nations 2014, 9). Perhaps, as some economists and political theorists have suggested, rich states could use immigration policy as a tool to address this problem. The principal proposal is that rich states lower their immigration restrictions to open up their labor markets to workers from poor countries. These workers would then be able to access higher paying jobs and remit some of their earnings to their families back home. Remittances are already an important source of income for poor countries; further immigration would enable even more to be sent (Barry 2011;Blake 2002, 280-2;Clemens and Bazzi 2008;Pritchett 2006). Not everyone is convinced by this proposal. Most political theorists writing on the ethics of immigration policy argue that migration is largely ineffective in relieving poverty and may sometimes be harmful (Brock 2009, 190-219;Miller 2005, 198-9;Pogge 1997;Wellman 2008, 128;Ypi 2008). Migration, they claim, tends to benefit better-off families who have the resources to send migrants abroad. The poor have little to gain. A further concern relates to skilled migration. The effect of international free movement, it is argued, is to deplete poor countries of much needed skilled labor. This suggests a second way by which immigration policy may be used to alleviate poverty. Rich receiving states could be encouraged to restrict skilled immigration in cases of brain drain, removing the main incentive skilled workers have to migrate (Ferracioli 2015;Kapur and McHale 2006, 317-8). Kieran Oberman is Chancellor's Fellow in Politics at University of Edinburgh, 22 George Square, Edinburgh EH8 9LF (kieran.oberman@ed.ac.uk).
I owe thanks to Steven Forde, the APSR editors and the three anonymous reviewers for their excellent comments and suggestions. Philip Cook, Sarah Fine, Daniel Halliday, and Lea Ypi provided invaluable feedback. I owe particular thanks to Daniel Butt for generously reading through multiple drafts and providing his characteristically acute observations on each. Finally, I am grateful to Edinburgh University's Department of Politics and International Relations for the support I received to conduct this research.
There are then two ways rich states could use immigration policy to alleviate global poverty: they could lift immigration restrictions if and when migration reduces poverty and impose restrictions if and when migration worsens poverty. The question whether immigration policy should be used in either of these ways is an important question for everyone who seeks an end to global poverty. While rich states do not, of course, decide their immigration policy on the basis of the needs of the global poor, there is significant disagreement within these states regarding how many migrants should be admitted and what the criteria for admission should be. Antipoverty campaigners must decide whether to intervene in immigration debates to back policies that would reduce global poverty or to put their limited time and energy into lobbying for alternative measures, such as higher development aid or the reform of international institutions.
There is both an empirical and a normative question to be addressed here. The empirical question is whether immigration policy can be an effective tool in countering poverty and, if so, what kind of policy would be effective: how open or restrictive to which sorts of migrants. The normative question is whether, if immigration policy can be effective, it should be used, given the importance of other values besides poverty alleviation. This article seeks to answer both these questions. It argues that the available empirical evidence suggests that a fairly open immigration policy, coupled with selective use of immigration restrictions in cases of deleterious brain drain, offers an effective tool to counter global poverty. Nevertheless, it holds that, for normative reasons, states should not use immigration policy to counter global poverty when alternative measures are available.
These conclusions contrast with those of most other political theorists writing on migration. The article finds little empirical support for the claim that migration and remittances fail to help the poor. In fact, the evidence suggests that migration benefits people across the income distribution. Even skilled migration has some positive effects such as creating networks, diffusing knowledge, and incentivising education. The only significant cases in which migration works against the interests of the global poor are those in which skilled migration leads to a net decrease in skilled labor that is not compensated for by other factors. There is every indication that a fairly open immigration policy, coupled with selective use of immigration restrictions in cases in which brain drain is a real problem, offers an effective means to reduce poverty. Empirically, there is nothing wrong with using immigration policy as one tool to address poverty.
The case this article makes against using immigration policy to address poverty is normative. Using immigration policy risks violating two human rights: the human right to stay in one's own country and the human right to immigrate to other countries. States violate the human right to stay if they fail to offer the global poor a means to achieve a minimally decent life in their own country. To the extent that states rely on migration to address global poverty, rather than assisting poor people in their own country, they leave poor people without any reasonable alternative to migrating. States violate the human right to immigrate if they impose unnecessary immigration restrictions. Immigration restrictions are unnecessary to counter brain drain if there are alternative solutions available.
The invocation of the human rights to stay and to immigrate marks a second point of contrast with most political theorists debating the ethics of immigration policy, for these rights have not been taken sufficiently seriously. While the importance of the human right to stay and its implications regarding poverty alleviation has been overlooked, the very existence of the human right to immigrate has been denied. This article defends these rights, showing how they are derivable from other rights already recognized in international law. It also explores the underlying foundations of these rights: the interests people have in pursuing the possibilities they wish to pursue and honoring the attachments they have made. Together, the human right to stay and the human right to immigrate entitle people to make their own migratory decisions. Recognizing these rights means recognizing that people in poor countries are not tools to be used for the task of poverty alleviation, but autonomous persons who should be allowed to decide for themselves whether or not to migrate.
Before embarking, let me make three points of clarification. First, while the article criticizes the use of migration to address global poverty, it is not against allowing poor people to migrate. On the contrary, it defends the idea of a human right to immigrate. Everyone, including the poor, has a right to cross international borders. What states cannot do is justify a failure to seek alternative means of addressing global poverty on the basis of having lifted immigration restrictions instead. Acting in this way risks leaving the poor without any reasonable alternative but to migrate. Our goal must be a world in which everyone is free to migrate, but no one is forced to migrate. Both lifting immigration restrictions and alleviating poverty are important; neither should be used as a substitute for the other.
Second, informed by how international law treats conventionally recognized human rights, the article interprets the human right to stay and the human right to immigrate as nonabsolute. The effect of this is that the article does not prohibit the use of immigration policy to address global poverty under all circumstances. The use of immigration policy can be justified if there are cases in which alternatives are ineffective. We should distinguish between the use of immigration policy as a policy choice, when it is used in place of alternatives that would also be effective in alleviating poverty and the use of immigration policy as a policy of last resort, when alternatives have proved ineffective. 1 This article condemns the former but not the latter.
Third, given that the article does not favor using migration to alleviate poverty, it is worth explaining why the article takes pains to show that migration is effective towards this end. The article seeks to encourage a shift in the debate within political theory so that it is better informed by relevant social scientific research and consequently has a better sense of the normative issues at stake. It is precisely because migration can be a powerful tool for alleviating poverty that the questions of whether it should be used, and, if so, under which conditions, assume such importance. As long as political theorists continue to dismiss migration as ineffective at alleviating poverty, they are unlikely to recognize that these are important normative questions that deserve their attention.
The empirical section of this article is thus an important precursor to the two normative sections that follow. Having established, in the first section, that a fairly open immigration policy is effective in alleviating global poverty, the rest of the article considers whether rich states should adopt this approach. The second section defends the existence and importance of the human rights to stay and to immigrate. It explains how both rights follow from the best interpretation of a set of human rights already recognized in international law. The third section explains why an appreciation for these rights cautions against the use of immigration policy to counter poverty. It contends that if people are to decide for themselves whether to stay in their home country or migrate abroad, rich states must seek alternative means to alleviate poverty. The conclusion the article reaches is that while immigration policy can be an effective tool to address global poverty, for normative reasons, alternatives must be sought.

THE EMPIRICAL DEBATE
Most political theorists believe that rich states have a duty to assist in the fight against global poverty. There is disagreement over how extensive this duty is. Some hold that rich states have a duty to eliminate most, if not all, arbitrary inequalities between their own citizens and those of poor states (Beitz 1999;Caney 2005;Moellendorf 2002). Others hold that global inequalities are not themselves morally troubling. On this second minimalist account, rich states are obligated merely to assist in the task of bringing the world's poorest up to a minimum threshold of wellbeing. Thereafter, each state should be left free to pursue its interests much as it sees fit (Blake 2001;Miller 2007;Rawls 1999).
This article remains neutral between these rival accounts of global distributive justice. All it assumes is that rich states have, at least, a weighty moral duty to assist in the fight against "desperate poverty." By desperate poverty I refer to a level of depravation so severe that those who suffer it are unable to fulfil their need for food, medical care, housing, or other basic goods. Desperate poverty prevents people from obtaining what John Rawls terms "minimum economic security" (Rawls 1999, 65), leaving them particularly vulnerable to injury, abuse, homelessness, and disease. The desperately poor may take considerable risks for the sake of small material gains (Sen 2000, 10;Wolff and De-Shalit 2007, 63-73). By depriving people of their human right to subsistence, moreover, desperate poverty frustrates the enjoyment of other human rights, whether it be the right to political participation, education, equality before the law or, as we shall see, the human right to stay in one's own country (Shue 1996, 22-34).
Of course desperate poverty cannot be sharply distinguished from poverty of less severity. Poverty is a spectrum; desperate poverty shades into poverty of less severe kinds. Partly for this reason it is impossible to specify precisely how many desperately poor people there are. What we can say is that, in the contemporary world, the desperately poor constitute a significant proportion of the human population. Given how little a person can purchase with $2 (PPP) a day, the 2.5 billion people living on less than this certainly fit within the category (Pogge 2013, 210). The desperately poor, moreover, can be found across the developing world, in middle-income countries such as China, Pakistan, and Nicaragua, as well as low-income countries such as Ethiopia and Malawi (Alkire et al. 2013;Sumner 2012).
This article focuses on desperate poverty because there is broad agreement that its alleviation is morally required. Those who adopt the minimalist account of global justice agree that rich states have a duty to alleviate poverty of this severity. Those who adopt an egalitarian account insist that justice requires much more than this, but do not deny that the alleviation of desperate poverty is a task of primary importance.
If political theorists are agreed that rich states have a duty to alleviate desperate poverty, how might this duty be fulfilled? Perhaps immigration policy could play a role. By loosening immigration restrictions, rich states could allow people from poor countries to access higher paying jobs and remit money to their families back home. The amount of money already being remitted is impressively high. In 2009, remittances totalled $307 billion, well over twice the amount developing countries received in foreign aid (World Bank 2011, 17). In many cases, remittances constitute more than 10% of GDP, and in some cases much more than that: 35.1% in the case of Tajikistan, 22.9% in the case of Nepal, 24.8% in the case of Lesotho (World Bank 2011, 21.) But does migration, and the remittance flows it generates, really translate into poverty reduction?
If one relied on the work of political theorists to answer this question, one would come to a depressing conclusion. Most theorists have painted migration as, at best, an ineffectual means to address poverty and, at worse, a dangerous force that exacerbates it. David Miller, for instance, contends that a policy of encouraging immigration . . . will do little to help the very poor, who are unlikely to have the resources to move to a richer country. Indeed, a policy of open migration may make such people worse off still if it allows doctors, engineers, and other professionals to move from economically undeveloped to economically developed societies in search of higher incomes, thereby depriving their countries of origin of vital skills (2005,198).
Citing this passage, Christopher Wellman argues that "the ardent egalitarian may not only be in no position to demand that affluent societies open their borders but she may also be forced to insist that they do not do so" (2008,128). Similarly, Thomas Pogge criticises antipoverty activists who put their time into lobbying for freer migration rather than structural reforms. According to Pogge, "not many of those that rich countries admit are really amongst the worst-off" but are rather "more privileged persons" with the resources to cross borders (1997,14). Pogge acknowledges that migrants send remittances, but he believes that these funds could only be a significant help to better off families. Some "funds may 'trickle down'" but since, Pogge contends, remittances are "more likely to increase than decrease domestic inequality" they are a "mixed blessing at best" (1997,15). Lea Ypi shares in the climate of pessimism, highlighting the issue of brain drain. Her primary example is the depletion of medical professionals from sub-Saharan African countries. She notes that these countries need more medical professionals to cope with extremely high rates of disease and possess scant resources to waste on training people who subsequently migrate abroad (2008,402). Ypi goes further, however, asserting the general principle that "[t]he higher the exit of a skilled workforce in a particular state, the more the welfare standards of the remaining citizens will be negatively affected" (2008,411). These thoughts lead her to seriously consider (although not ultimately endorse) the idea that a utopian world might be a world of closed borders.
Finally, Gillian Brock claims that immigration, in the absence of other antipoverty measures, risks "mixed results, and could even constitute a considerable step backward, for global justice" (2009,191). Like Pogge, she contends that migrants are from better off families and thus remittances rarely reach the poor (2009,205). One might still hope that remittances have an indirect effect upon poverty by spurring the economy, but that too appears unlikely to Brock since, according to her, the vast bulk of the money (she cites a figure of 90%) is spent on "consumer goods and daily living" rather than "savings and investment" (2009,206). These supposed problems with remittances, coupled with brain drain, make immigration a poor policy choice for any government interested in addressing global poverty. 2 The picture these theorists paint of the relationship between migration and development is a gloomy one. But how seriously should we take their reports? One warning sign lies in the limited use of empirical sources. Neither Miller, nor Wellman, nor Pogge cite a single source to support the claims that the poorest people tend not to migrate, that brain drain is straightforwardly deleterious and that remittances create inequality. Ypi and Brock do cite some empirical studies, but they too fail to address the by now extensive literature suggesting more positive conclusions.
When one turns to the empirical literature, the first thing that stands out is the sheer quantity of research affirming the effectiveness of migration in reducing poverty. Studies on migration from Latin America (Acosta et al. 2008 (Pernia 2008), Zimbabwe (Bracking and Sachinkonye 2010), and Tonga and Fiji (Brown and Jimenez 2008), all report that migration and remittances reduce poverty. One prominent global study, by World Bank economists Robert Adams and John Page, finds that a 10% increase in the share of emigrants in a country's population leads to a 2.1% decline in the number of people living on less than $1 a day. It also finds that a 10% increase in the amount of remittance money a country receives leads to a 3.5% decrease in the poverty rate (Adams Jr and Page 2005). Increased income is, moreover, only one of several ways by which remittances benefit the poor. Remittances have been shown to be associated with lower rates of infant mortality and underweight newborns (Kanaiaupuni and Donato 1999;McKenzie 2005, 128). There is also evidence that remittances increase school attendance and literacy rates and decrease child labor (López-Córdova et al. 2005;Yang 2008). It is, in fact, difficult to find any recent study on remittances that does not regard them as an effective means to reduce poverty. 3 But how can migration reduce poverty if, as a number of theorists have claimed, it is mostly the better off who migrate? And is Pogge not right to contend that remittances create inequality and are thus a "mixed blessing at best"? In fact, both claims are controversial. While some studies report that migrants are primarily from less poor families and that remittances 2 While Brock's main complaint against migration is its supposed ineffectiveness in addressing poverty, she does suggest some additional normative concerns. For instance, she contends that high rates of migration can sap the social and economic life within a sending state, leaving people with no option but to migrate in order to find work (Brock 2009, 205-6). Insofar as Brock objects to migration on other grounds, besides its supposed ineffectiveness in addressing poverty, I make no criticism of her argument. Indeed, her concern with forced migration is one I share and develop below. 3 There was an early period of research that was much less optimistic. See for instance Lipton (1980). This earlier research may have established certain "migration myths" that have, in turn, affected political theorists writing on this subject. I return to this point below.
increase inequality (Adams Jr 1989;Wouterse 2010), others present contrary findings (Acosta et al. 2008;Taylor and Wyatt 1996). One explanation for these divergent results is that the character of migration changes during the course of a community's migration history. When people from a particular community first start migrating, the costs of migration are high and it is, generally, less poor families from which pioneers are drawn. As time passes and social networks develop, the costs of migration diminish and larger numbers of poorer people migrate. Eventually a stage is reached in which poorer families receive the largest share of remittances, thereby decreasing inequality. This explanation was first proposed by Oded Stark, J. Edward Taylor, and Shlomo Yitzhaki (1986). It has since been affirmed by numerous subsequent studies (Koechlin and Leon 2007;McKenzie and Rapoport 2007;Taylor et al. 2005). The lesson for rich states seems to be that they should accept not only more migrants but large numbers of migrants from particular communities in order to encourage the development of social networks that make it easier for poorer people to migrate.
A second point worth making here is that people can benefit from remittances even if they do not receive them. Since remittances are often spent on local consumption and investment, they can provide a significant boost to production, employment, and growth. Economist J. Edward Taylor estimates that for every dollar remitted, Mexico's GNP increased by somewhere between $2.69 and $3.16. He concludes that "[d]espite the far reaching impact that migration and remittances can have on households that send migrants, it appears that most of migration's effects are found outside these households" (2006,208). The primary beneficiaries of these indirect effects, moreover, are "relatively poor and middle-income rural and urban families" (1999, 70). Brock's suggestion that remittances are spent predominantly on consumer goods and daily living is controversial. Migration specialist Hein de Haas regards the idea as a "migration myth"-a holdover from early days of remittance research that has taken time to die despite mounting evidence to the contrary (2005,1274). The 90% figure that Brock cites is almost certainly wrong. 4 Most studies report a much lower figure. Economist Robert Adams, for instance, reports that a mere 12% of remittances in rural Egypt are spent on daily consumption (Adams Jr 1991, 720). Other studies find significant amounts being invested in education, agriculture, and businesses (see .
None of this is to suggest that migration is a panacea. Even if there were not the normative objections to using migration to counter poverty developed in the proceeding sections, other measures would still need to be taken. However widespread the benefits of migration are felt, there will, undoubtedly, be some people who miss out. For this empirical reason, migration cannot be the sole means by which we seek to eradicate global poverty. But this point is entirely compatible with the claim that migration is one effective means to reduce poverty.
Let us turn to the issue of brain drain. Here there is a matter of genuine concern. Brain drain does seem to be a significant problem for some sectors in some countries. Medical brain drain from sub-Saharan Africa is understandably the most commonly cited example given the region's low doctor to patient ratios, constrained health budgets, and high rates of AIDs/HIV. Nevertheless, a wholly pessimistic view of skilled migration is misleading. Ypi's claim that "[t]he higher the exit of a skilled workforce in a particular state, the more the welfare standards of the remaining citizens will be negatively affected" is highly questionable.
Skilled migration has a number of important benefits: it generates remittances (Bollard et al. 2011), establishes social networks (Meyer 2001), and diffuses knowledge (Le 2008). Perhaps the most important positive effect, however, is incentivising education. The possibility of earning large salaries abroad leads people to invest in education. If more people acquire skills in the hope of migrating than there are opportunities abroad, then a poor country will actually experience a brain gain, achieving higher numbers of skilled workers than if the borders had been closed (Stark 2004). The extent to which this positive result obtains is the subject of on-going debate. One prominent study finds that skilled migration generates a 3% increase in the number of skilled professionals living in the developing world, but also notes that not all countries experience a brain gain and that small states are much more likely to be losers than winners (Beine et al. 2008). Skilled emigration, it seems, is a problem for some but not all countries.
Where brain drain is a problem, there may be measures poor states can take to ameliorate it. They could, for instance, look to raise salaries and improve working conditions to dissuade skilled workers from leaving (Kapur and McHale 2005, 197-9). Where other solutions fail, emigration restrictions might be an option. While the human right to emigrate is recognized in international law, restrictions are permitted when necessary to prevent the kind of grave cost that deleterious brain drain constitutes (see the following section below).
There is reason, however, to think that poor states cannot solve the problem of deleterious brain drain alone. The gap between rich and poor states is such that as long as skilled workers have the opportunity to migrate to rich states, they have a powerful incentive to leave. Poor states can try to compete, but, with limited budgets, they will struggle to match the pay and conditions obtainable in rich states (Connel 2010, 177). Even emigration restrictions offer no assured solution. It is striking how rarely emigration restrictions are now imposed, even by states suffering severe brain drain. States that may have wanted to impose emigration restrictions have been defeated by a combination of domestic political opposition and the practical difficulty of enforcement (Fitzgerald 2006;Haas and Vezzoli 2011). As long as poor states are unable or unwilling to solve the problem of deleterious brain drain themselves, there is a case for rich states to take action. Reducing admission of skilled workers from countries and sectors suffering deleterious brain drain is one action rich states could take (Kapur and McHale 2006, 317-8).
This section has investigated the empirical question of whether migration forms an effective means to address global poverty. The answer arrived at is broadly "yes." Remittances help poor people escape poverty, benefiting not only the people to whom they are sent but also people in the surrounding community. Remittances raise the incomes of the poor, lower infant mortality and increase literacy and school attendance. Even skilled migration can yield benefits for those left behind. The only cases in which migration does seem to harm the poor are those in which the benefits of skilled migration fail to cover the costs. Deleterious brain drain is a real, if sometimes exaggerated, problem. 5 It would seem then that rich states could use immigration policy as one means to fulfill their duty to assist in the alleviation of global poverty. An effective counterpoverty immigration policy would involve lifting restrictions against most migrants but imposing restrictions against skilled workers from countries and sectors suffering deleterious brain drain. This combined package of increased immigration with selective restrictions could yield a significant reduction in global poverty. 6 But should rich states adopt this approach?
In what follows, I shall argue that rich states should not use immigration policy to address poverty when effective alternatives are available. Doing so risks violating two human rights: the human right to stay in one's state and the human right to immigrate to other states. Further below, I explain how these rights can be violated by the use of immigration policy to counter poverty and consider what alternatives may be available. First, however, let me defend the existence and the importance of the rights in question.

THE RIGHT TO IMMIGRATE AND THE RIGHT TO STAY
I shall start with the more controversial of the two: the human right to immigrate. The argument for this right follows from the most plausible interpretation of a set of human rights already recognized in international law. This set I shall term "human freedom rights." The set includes the right to freedom of association, expression, religion, occupational choice, marriage, and a subset of rights relating to migration. Human freedom rights allow us to make certain basic life choices including the choice of which (if any) religion we practice, whom we associate and communicate with, whom (if anyone) we marry, and which career we pursue. Human freedom rights entitle us to make these choices on our own terms, without government restrictions on our range of options. Under ordinary circumstances, governments should not prevent us from joining the associations we wish to join, attending the public meetings or cultural events we wish to attend, practicing a particular religion, pursuing a given career, or marrying the consenting partner of our choice. Our range of what we might term "life options"-friends, family, civic associations, expressive opportunities, jobs, and marriage partners-should not be subject to government restriction.
If this is what human freedom rights are, what about the subset of human freedom rights relating to migration? Migration related rights are listed in a number of international human rights documents. Article 13 of the Universal Declaration of Human Rights (UDHR 1948), for instance, holds that (1) Everyone has the right to freedom of movement and residence within the borders of each state.
(2) Everyone has the right to leave any country, including his own, and to return to his country.
These rights play an essential role in securing the free exercise of all other human freedom rights. If one is banned from a certain geographical area, such as a province within a country or a country itself, then one cannot visit friends or family, attend a religious or educational institution, express one's ideas at a meeting or cultural event, seek employment, or pursue a love affair, anywhere within that area. One's right to freedom of movement is curtailed, but so are one's rights to freedom of association, expression, religion, occupational choice, and marriage.
If human freedom rights protect our access to an unrestricted range of life options, should their scope not be much broader than how they are conventionally defined? In fact, human freedom rights, as defined in law, are extremely broad. Article 13(1) of the UDHR, and its equivalent in the International Covenant on Civil and Political Rights (ICCPR 1966, Article 12(1)), entitle people to move freely across the entire territory of a country, no matter its size. Moreover, these articles entitle everyone who is legally present within a state to enjoy free movement, foreigners as well as citizens. The same rule applies in the case of most other human freedom rights: legally present foreigners are equally entitled to freedom of association, expression, religion, and marital choice (Human Rights Committee 1986). Nor are migration restrictions beyond the purview of international human rights law. As is clear from Article 13(2) of the UDHR, the use of unjustified emigration restrictions against any person, as well as the use of unjustified immigration restrictions against citizens, violates legal human rights.
In sum, human freedom rights, conventionally defined, demand that governments allow us to (1) access all life options within our own state, (2) leave our own country in search of life options elsewhere, (3) access most of the life options available within foreign states that choose to admit us, and (4) return to our own state whenever we so choose. As we move about the world, exercising our human freedom rights, the only kinds of government restriction that remain beyond the purview of human rights law are immigration and employment restrictions imposed against foreigners. These two kinds of restriction are, moreover, closely connected. Since people ordinarily need to work in order to live, restrictions on employment are effectively restrictions on immigration as well. Reflecting this fact, I shall henceforth use "immigration restrictions" broadly to refer to all government restrictions that, directly or indirectly, prevent foreigners from living within a country. Now when noting the extensive nature of conventional human freedom rights, we should acknowledge that these rights are also internally bounded and nonabsolute. They are internally bounded in the sense that one person's life options are limited by other people's choices. Were someone to force you to marry him, adhere to his religion, or join his private association, he would not be exercising his rights to freedom of marriage, religion, and association but violating yours. In this regard, the limitation that you would place on this other person's options is quite unlike government restrictions on the marital, religious, and associational choices of consenting adults. The first kind of limitation results from an autonomous choice; the latter constitutes its very denial (cf. Miller forthcoming; Wellman 2008).
Human freedom rights are nonabsolute in the sense that governments can justifiably restrict human freedom rights when restrictions are necessary to prevent grave costs to security, subsistence, public health, and other matters of primary importance. Thus a government might legitimately ban a demonstration that would result in widespread rioting but not to prevent some more minor cost, such as offense to the demonstrator's opponents or temporary traffic disruption. Human rights are important but nonabsolute; they are "resistant to trade-offs but not too resistant" (Griffin 2008, 37). This fact is already recognized in international human rights documents, such as the UDHR and the ICCPR, which list reasons that justify curtailment. 7 This then is how human freedom rights are conventionally understood. They are extensive (entitling foreigners and citizens to make their own life decisions), internally bounded (each person's range of choice is limited by the free choices of others), and nonabsolute (curtailment can be justified when necessary to avert 7 See the UDHR, Article 29(2), and the ICCPR, Articles 12(3), 18(3), 19(3), and 22(2). grave costs). But once we understand human freedom rights in this conventional way, we immediately see how anomalous immigration restrictions are. For immigration restrictions represent a severe restriction on the freedom of both foreigners and citizens. They restrict the freedom of foreigners since they deny excluded foreigners access to civic associations, religious institutions, friends, family, romantic partners, jobs, educational opportunities, and opportunities for expression (protests, public meetings, conferences, etc.). They restrict the freedom of citizens since they deny citizens the choice to marry, befriend, associate with, learn from, worship with, or gain employment from excluded foreigners. Such restrictions cannot be justified as the limit set upon each person's options by the life choices of others. Were immigration restrictions lifted, citizens could still refuse to marry foreigners, worship with foreigners, or join their private associations. Personal decisions, such as these, would be left to individuals, not foisted upon them by government. 8 The only solution to this anomaly is to recognize a human right to immigrate alongside those human freedom rights already recognized in international law. Like these other rights, the human right to immigrate would be nonabsolute. Immigration restrictions could be justified if they were necessary to prevent some grave cost to security, subsistence, public health, etc., but not for the sake of more minor goals. In other words, we should award the human right to immigrate exactly the same weight that we award the other human freedom rights from which it is derived.
Consider, in this regard, a common argument for immigration restrictions: that they defend the economic interests of poor citizens. Immigration, it is sometimes said, increases competition at the bottom of the labor market, driving down wages and driving up unemployment. This claim is in fact highly controversial (compare Borjas 1999;Card 2005), but let us suppose, for the sake of argument, that it is true. If immigration harms the interests of poor citizens, can restrictions be justified? Following what has been said above, we need to distinguish between different levels of harm.
At one extreme, we can imagine a scenario in which unrestricted immigration drives the poorest citizens of the receiving state into desperate poverty. Poverty of such severity certainly constitutes a grave cost and could justify restrictions on human freedom rights, including the human right to immigrate, were there no less coercive means to address it. But now suppose, by contrast, that the effect of immigration would simply be to lower the position of the poorest citizens relative to other citizens while leaving them with enough to fulfill basic needs for food, medical care, housing, and so forth. Such a situation is troubling, but it cannot justify restrictions of a human right. The example of other human freedom rights makes this plain, for their exercise too can have deleterious distributive implications. To give just two examples: freedom of marital choice allows wealthy people to marry each other, frustrating one means by which poor people can better themselves, and freedom of occupational choice enables talented people to bargain for a share of the social product that could have gone to the poorest instead. Yet in neither of these cases would restrictions be justified (Cohen 2008, 118-222;Wellman 2008, 26). The consensus position remains the Rawlsian one: fundamental liberties should not be restricted for the sake of achieving improvements in the distribution of people's material resources once basic needs have been secured (Rawls 1973, 476).
In taking this position, we need not be indifferent to the claims of poor citizens. Governments have a range of tools at their disposal to address domestic poverty. They can, for instance, impose redistributive taxation, expand educational opportunities, and subsidize social housing. Governments cannot use the existence of human freedom rights as an excuse for failing to address domestic poverty by other means.
Similarly, the argument for a human right to immigrate does not condemn all distinctions between citizens and foreigners. Governments should not prevent foreigners from living within their territory and exercising their other human freedom rights, 9 but this does not mean that foreigners must be offered every benefit that citizens enjoy. One can imagine a world in which people are free to live and work where they choose, but are prevented from accessing certain privileges (voting rights, higher education grants, small business loans, etc.) in countries other than their own. Such a world would in fact resemble the contemporary United States and the European Union, both of which allow citizens to migrate freely from one member state to another, but also allow state governments to reserve certain benefits for their own residents (van der Mei 2002). Such discrimination may be just or unjust. The human right to immigrate does not settle the matter either way.
The idea of a human right to immigrate is not then as demanding as it may first appear. It does not rule out restrictions when they are necessary to avert grave costs, nor condemn every distinction between citizens and foreigners. Nevertheless it denies states broad discretion to restrict immigration as they choose and that remains a radical proposition. Some will want to resist it by looking for ways to reinterpret our conventional human freedom rights to render them compatible with state discretion over immigration. One suggestion is to interpret conventional human freedom rights as entitling access to no more than an "adequate" range of life options: a reasonable choice of jobs, religions, friends, marriage partners, civic associations, and so forth (Miller 2007, 207;Pevnick 2011, 84-5). With only an "adequate" range accessible to them, people may not be able to live exactly the life they desire, but they can still exercise a degree of autonomy. That degree of autonomy, moreover, can be provided within each state. People thus have no essential interest in migrating abroad (Miller 2007, 207;Pevnick 2011, 85).
There are various problems with this suggestion, but the most striking is that it cannot ground conventional human freedom rights. If the smallest states offer an "adequate" range of options internally, larger states must offer more than an "adequate" range. On the "adequate" range view then, larger states could prohibit all but an "adequate" range of options internally without violating their citizens' human rights. They could ban disfavored books, civic associations, and religions, prevent people from leaving their subnational region or choosing their careers, and even deny people contact with their family, while still claiming to respect human rights. Yet our human freedom rights, conventionally understood, do not permit such severe government restrictions. Governments that would restrict our range of options to an "adequate" range by banning books, civic associations, religions, and so forth violate our human freedom rights.
Still, it is worth asking why our human freedom rights are this extensive. What essential interest do people have in accessing options beyond an "adequate" range? To answer this question, it is helpful to introduce a distinction between two kinds of life options: "attachments" and "possibilities." "Attachments" are those life options that we have chosen or in some other way become attached to, such as our family, friends, career, and religion. "Possibilities" are life options that we are not currently attached to but may wish to pursue in the future. 10 It is clear enough that people have an essential interest in accessing attachments, whether or not their attachments fall within a supposedly adequate range. For instance, everyone has an essential interest in being with the people they love and religious people have an essential interest in practicing their religion. If someone's loved ones or religion falls outside of the supposedly adequate range accessible to her, she has an essential interest in accessing options beyond that "adequate" range. The fact that there are other people with whom she could form relationships, or other religions she could practice, changes nothing in this respect.
But what about possibilities? What essential interest do people have in accessing options beyond an "adequate" range to which they are not, as yet, attached? That our human freedom rights protect our ability to access possibilities, as well as attachments, is clear, for otherwise we could not explain the extensive scope of human freedom rights which permit us, among other things, to join and establish new civic associations, meet new people and make new friends, learn about, and convert to, any religion we choose, attend meetings on subjects we know little about, and move to areas to which we have little, or no, connection. Explaining precisely why this is the case is a complicated matter. I have sought to address it elsewhere (Oberman forthcoming). Here, let me present a general thought that lies behind the unrestricted range view. The thought is 10 Joseph Raz (1986, 411) makes a similar distinction. this: the primary purpose of human freedom rights is not, in fact, to ensure that individuals have a set number of options to choose from, but rather to isolate from the host of issues over which governments rightfully exercise power a subset of matters that each individual should be allowed to determine for herself. These matters include where she lives, whom she lives with, who her friends are, whom she marries, which religion she practices, which associations she joins, what work she does, and how she spends her free time. When governments interfere in these matters, without strong justification, they overstep their authority. They overstep their authority when they use police and internal checkpoints to deny people certain options and they overstep their authority when they use border fences and deportation squads. Immigration restrictions, like internal restrictions, trespass upon the personal domain.
In defending people's rights to access both attachments and possibilities, I do not deny that there is, standardly, an important moral difference between the two. In most cases, it is more important for people to be able to access options to which they are already attached than mere possibilities. A person has a particularly strong interest in being with her family, pursuing her career, practicing her religion, and taking an active part in her community. So more can be expected of governments to enable people to honor their attachments than to enable people to pursue possibilities. If a person from London wants to start a new life in Edinburgh or Chicago, but has no attachments in either place, the main thing that is demanded of the governments involved is that they do not prevent her from moving. It is not clear that they have positive duties to help her move by, for instance, paying her fare. Things seem different, however, when people have attachments in an area and particularly when almost all of their attachments lie within one area. In such cases, there is an argument for government action to ensure that people can maintain their attachments.
These thoughts bring us to the second right I wish to defend: the right to stay in one's own state. If human freedom rights are interpreted as entitling people to access an unrestricted range of life options, then clearly people have a human right to stay. If they are forced to leave their state, they cannot access options within it. But additionally, for most people, the options that represent their most important attachments are situated within their own state. Thus, for most people, the human right to stay is a particularly important right, more important than the human right to immigrate.
The human right to stay protects people against standard threats that could otherwise force them to leave their state. Someone is forced to leave if she leaves for want of a reasonable alternative. There are a variety of reasons why someone may lack a reasonable alternative to leaving. She may have no alternative, reasonable or otherwise, as when a person is tied up and driven over the border. More often, however, people are forced to leave because the alternative they have is one they cannot reasonably be expected to bear. This is true of people who comply with an order to leave having been threatened with death or assault if they remain. It is also true of people who must leave in order to escape those who actively seek to kill or assault them. Of the former group, we say they are expelled; of the latter that they flee persecution. 11 Expulsion and persecution are both standard threats to the right to stay. But another standard threat is desperate poverty. A life led in want of food, clean water, medical care, and other basic goods is not a reasonable alternative to the satisfaction of one's needs. Desperately poor people who migrate abroad in search of a better life are people who are forced to leave their state. Like the expelled and the persecuted, they have no reasonable alternative to migration.
Since desperate poverty is a standard threat to the right to stay, rich states, as well as poor states, have duties to address it. This claim is not a radical departure from the conventional view that rich states owe a duty of assistance to the global poor. The right to stay provides an additional reason for rich states to take action to alleviate global poverty, but they are obligated to do so in any case. The right to stay, nevertheless, has important implications for how rich states seek to alleviate global poverty. Since desperately poor people have a right to stay in their own country, rich states should seek measures that bring assistance and opportunities to the poor rather than relying on them to migrate. I develop this point in the next section.
In recognizing desperate poverty as a standard threat to the right to stay, we acknowledge that the right to stay is, in part, a positive right, entailing positive duties upon states to ensure that people can fulfill their basic needs within their home country. That the right to stay can ground these positive duties is a consequence of its importance. People who are forced to leave their country are forced to leave behind many of the people they love and projects they are engaged in. Forced migration tears at the fabric of people's lives. In that context, it seems right that states should be prepared to invest resources to ensure that people are not forced to migrate.
In recognizing that the right to stay is, in part, a positive right, we should not, however, overlook the fact that it also entails negative duties. States have duties not to force people to migrate by expelling, persecuting, or impoverishing them. On some accounts of the causes of desperate poverty, harmful policies adopted by rich states, and the international institutions that rich states support, play a crucial role. Rich states are accused of enacting unfair terms of trade, subsidising rich world farmers to undercut the poor, buying natural resources from illegitimate governments, and enforcing an intellectual property regime that prevents poor people from accessing life-saving drugs (Bacon 2013;Pogge 11 Of course, there will always be people who choose not to leave their state even when staying constitutes an unreasonable alternative, such as those who choose death or assault over exile. But this does not mean that those who do leave act voluntarily. The mere presence of an alternative is insufficient to render an action voluntary. Mugging victims have an alternative-to endure the violence that the mugger threatens them with-but clearly mugging victims do not hand over their wallets voluntarily. See Olsaretti (1998Olsaretti ( ), especially p. 71, n. 34. 2002Olsaretti ( , 2007. If these accounts are correct, rich states, in causing desperate poverty, create forced migration. The first duty of rich states would therefore be to stop enacting these policies, which would in turn help poor people to secure their human right to stay. Yet not everyone believes that rich state policies are the principal cause of desperate poverty. Some hold that corruption and misgovernment in poor countries are the most important causal factors (Landes 1998;Rodrik et al. 2004). If this alternative view were correct, would rich states retain a duty to assist people in poor countries? There is a normative issue here but there is also a practical one and it is worth separating the two. The normative issue is whether states can have duties to uphold human rights when other states are responsible for their violation. The prevalent view in the literature is that states can indeed have duties to act in such cases. The fact that it is state A that is responsible for the violation of its citizens' human rights does not, in itself, permit states B and C to stand idly by. 12 The practical issue remains, however, which is whether there is much rich states can do to address poverty in countries that suffer from corrupt or abusive governments. Perhaps the only recourse rich states have is to admit more migrants from these countries. I address this issue below.

THE CASE AGAINST USING IMMIGRATION POLICY TO COUNTER POVERTY
The first section of this article found that there is nothing wrong empirically with the idea of using immigration policy to alleviate poverty. The problem with this approach is not that it is empirically flawed but that it risks violating the two rights we have explored in the previous section. Using immigration as an alternative to directly addressing poverty in poor countries risks violating the human right to stay. Imposing immigration restrictions against skilled workers, when there are alternative means to address deleterious brain drain, risks violating the human right to immigrate. Let me develop each of these points in turn.
Since the human right to stay is an especially important right, states are required to search for ways to address desperate poverty in situ, so that desperately poor people can fulfill their basic needs within their home country. When rich states rely on migration to address poverty, instead of searching for alternatives, they fail in this regard. Poor people should not be forced to migrate hundreds of miles to fulfill their basic needs if alternatives are available.
But by what other means, besides immigration, can rich states address poverty abroad? Speaking broadly, there are two main proposals: (i) rich states could spend more on foreign aid, and (ii) rich states could make changes to global institutions, including rules governing trade, debt, resource ownership, and intellectual property rights. These proposals generate their own empir-ical debates. Aid critics point to the trillions already spent without clear evidence of success. They note the connection between aid and corruption and the seeming inability of donors to break it (Easterly 2006;Moyo 2009). Regarding the latter proposal, the debate is not so much whether rich states should make reforms but rather whether these reforms, if undertaken, would yield extensive results. Pogge, a prominent advocate of institutional reforms, claims that they could, at relatively little cost, eradicate global poverty (2007,30). However, since Pogge (2010, 176) also concedes that there has been insufficient empirical research on the matter, many will find it hard to share Pogge's confidence that the types of reforms he proposes will yield such dramatic results (Cohen 2010).
Underlying the debate over the effectiveness of aid and global institutional reforms is disagreement over the causes of global poverty. If global poverty is largely caused by domestic factors, such as corruption, then it seems unlikely that aid and global institutional reforms will do much to reduce poverty. Indeed, if it is domestic factors that cause poverty, it is unclear what rich states can do to affect change. They could try to exploit their position as donors by making aid to poor states conditional upon government reforms. But aid conditionality has been tried in the past, with questionable results (Dijkstra 2002). More coercive measures, including, at the extreme, military intervention, are not only morally and legally controversial but also fail to guarantee success. As intervention in Afghanistan, Iraq, and Libya has demonstrated, it is easier to bring down corrupt and abusive governments than to put honest and effective governments in their place.
Perhaps then, a defense could be made of the use of immigration to address poverty, if one assumes a sufficiently pessimistic view of the available alternatives. As I have argued from the outset, the use of immigration to address global poverty can be justified if alternative measures prove ineffective. However, I would like to enter a hypothesis here that there is at least one alternative that should prove equally as effective in combatting poverty as migration. The alternative is this: rich states could act much like remittance senders by sending cash transfer payments to poor people abroad. There seems no reason to think that cash transfers sent by rich states should be any less effective in combatting poverty as cash transfers sent by individual migrants. What is important for achieving a reduction in poverty from cash transfers is that the recipients receive the cash; the identity of the sender does not seem significant.
Of course, implementing this alternative requires the agreement of the recipients' governments. It would also require some level of bureaucracy to distribute the funds to those entitled to receive them. These are not demanding requirements, however, when compared with traditional forms of aid. Traditional forms of aid involve complex operations in which money changes hands many times, offering multiple opportunities for corrupt officials to pocket funds or obtain kickbacks from suppliers. With cash transfer programs, the route from donor to recipient is much more direct. Indeed, with no procurement involved, kickbacks are unobtainable (Chene and Transparency International 2010, 5;Kapur et al. 2008, 40). Cash transfers, like remittances, bypass middlemen, allowing the poor to decide how resources are spent.
The idea of using cash transfers to address poverty is an emerging area in development economics. Cash transfer programs have been initiated in South Africa, Mexico, Brazil, and elsewhere, with some impressive results (Hanlon et al. 2010). Cash transfers have been shown to significantly boost consumption of proteinrich food with consequent benefits for child nutrition (Aguero et al. 2006;Attanasio et al. 2005). Children from families receiving cash transfers tend to be taller and less susceptible to anaemia (Gertler 2004). Even cognitive improvements in memory and language acquisition have been reported (Paxson and Schady 2010). Like remittances, cash transfers are not only spent on consumption. One study of the Mexican Oportunidades program, for instance, found that for every peso a recipient spent, 26 cents were invested, increasing agricultural income by 10 percent after 18 months (Gertler et al. 2012). Also like remittances, there is evidence that cash transfers yield benefits for nonrecipients (Angelucci and De Giorgi 2009).
While cash transfer programs have tended to be funded by national governments, there is no reason why funding could not come from rich states. Such programs are in their relative infancy and there is much research yet to be done. But, given the promising results and given the similarities between cash transfers and remittances, it seems plausible that cash transfer programs could prove an effective alternative to migration in many cases. Whenever cash transfer programs provide an effective alternative, they are in one important respect morally preferable: they do not require desperately poor people to migrate abroad.
The argument I have made against using immigration to address poverty has concentrated on the human right to stay. Before moving on, however, it is worth noting that a different argument can be made from distributive justice. There are various ways to understand the idea of a duty to assist on the part of rich states, but one plausible way is that rich states should share, along with the poor, some of the burden of relieving poverty. When poor people migrate long distances and sacrifice much of their income for the sake of family members, it is clear they are accepting a considerable burden. It is much less clear that rich states that admit them are sharing that burden, especially since rich states typically benefit economically from immigrant labor (Borjas 1999, 87;Felbermayr et al. 2010). From this perspective, using immigration to address global poverty appears less like fulfilling a duty to assist and more like exploiting the people to whom the duty is owed. If both rich states and poor people have duties to alleviate poverty, why should the latter bear a disproportionate share of the costs?
We have established that immigration should not be used to address poverty when there are effective alternatives available. Let us turn to the idea of using immigration restrictions to address brain drain. Here too alternatives must be sought. Skilled workers have a human right to immigrate, like everyone else. The human right to immigrate should not be curtailed unless doing so is necessary to avert a grave cost. Deleterious brain drain is indeed a grave cost, but it is not obvious that immigration restrictions are necessary to avert it. Alternative measures include raising salaries and improving working conditions of skilled workers in poor states, as well expanding higher education so that more skilled workers can be trained. As noted, poor states may lack sufficient resources to successfully pursue these measures. But if rich states were to provide significant supplementary funds, a sufficient number of skilled workers might be trained and retained without any recourse to migration restrictions. 13 In making this argument, I do not deny that skilled workers could have a duty to stay and assist their poor compatriots (cf. Tesón 2008). Perhaps some skilled workers do have a duty to stay, partly because they should repay the costs of their training and partly because everyone, who is in a position to help, has a duty to assist desperately poor people. 14 The argument, rather, is that no one should be prevented from exercising a human right, even in contravention of a moral duty, unless it is necessary to avoid some grave cost.
It is again helpful to consider analogous examples involving other human freedom rights. Arguably, skilled workers in rich states, such as doctors and politics professors, have moral duties not to threaten to change occupation or to emigrate abroad, simply to secure higher wages. But it would be wrong, indeed a human rights violation, to prevent these skilled workers from changing profession or emigrating abroad. Things might be different if rich states were so short of resources that they could not afford to pay the higher salaries. The threat of unmanned hospitals and closed universities might provide grounds for restricting human rights in various ways. But since rich states have sufficient resources to ensure that sufficient numbers of skilled workers are retained, restrictions are unjustified. The distribution that results may be unfair but that unfairness must be tolerated. Not all moral duties can be enforced. Sometimes people have a right to decide whether or not they fulfill a moral duty incumbent upon them. In this sense, people may be said to have a right to violate their duty (Enoch 2002;Waldron 1981). Doctors and professors in rich states have a right to violate their duty not to migrate or change profession. 15 Exactly the same is true of skilled workers from poor states. Skilled workers may have a duty to stay, but their human rights cannot be restricted to avert a cost that can be averted by some more acceptable means. Rich states are thus not justified in restricting the human right of skilled workers to immigrate unless these states are unable to address deleterious brain drain by other means, such as providing the funding to raise salaries, improve conditions, or train replacement workers.
Before concluding, let me consider one last objection. The article has argued that the use of immigration policy as a means to address global poverty is often incompatible with the human right to stay and human right to immigrate. I have conceded, however, that rich states might be justified in using immigration policy to address poverty when other measures prove ineffective or generate grave costs of other sorts. This concession might seem a betrayal of the principle the article has sought to defend. If it is so important that people living in poor sending states have the freedom to choose whether or not they can migrate, why should it ever be permissible for states to deny them this freedom?
The answer to this question is not that the freedom to choose whether or not to migrate ever ceases to be important, but rather that there are circumstances in which upholding this freedom proves incompatible with tasks of even greater moral importance. The task of eradicating desperate poverty is one of particular moral urgency. Awful as it must be for a person to be prevented from migrating to a state she wishes to enter or be forced to leave her country to escape desperate poverty, it is even worse for a person to live in desperate poverty without any means of escape. The human right to basic subsistence must then take precedence. Respecting the human right to stay and the human right to immigrate does not mean denying the possibility of justifiable restrictions, but rather insisting that restrictions are never imposed except when necessary to avert grave costs.

CONCLUSION
The available empirical evidence suggests that migration is effective at reducing poverty. The pessimistic claims of political theorists are largely ungrounded. There is nothing wrong empirically with the proposal to use immigration policy as one tool to counter poverty. The problem with this proposal, we have found, is normative. People should be free, when making important life decisions, to pursue new possibilities and honor the attachments they have already made. For these reasons, people have a human right to immigrate to other states and a human right to stay in their own. Unless these rights are respected, people's life choices are significantly constrained.
Since people have a human right to immigrate, immigration restrictions should not be imposed when there are alternative means to counter deleterious brain drain. Since people have a human right to stay and since this right is of particular importance, people are entitled to assistance in fulfilling their basic needs without the requirement that they migrate abroad. When there are no alternatives available in either case, immigration policy can be used to address poverty. But immigration policy should be regarded as a policy of last resort. When states employ it as a policy of first resort they risk unnecessarily curtailing the freedom of people in poor countries to live their lives as they choose. Poverty is reduced but at an unacceptably high price.