Sufficiency, Priority, and Selecting Refugees

Journal of Applied Philosophy doi: 10.1111/japp.12445 Sufficiency, Priority, and Selecting Refugees MOLLIE GERVER There are more refugees than many states have a duty to accept. As a result, many states are permitted to accept only some refugees and not others. This raises the question of how states should select these refugees. I defend the claim that states should resettle the worst-off refugees, and maximize the number of refugees gaining a sufficiently decent life. When resettling the worst off conflicts with maximising the number gaining a sufficiently decent life, states should select refugees to maximise aggregate benefits, with benefits holding greater weight the further refugees are from a sufficiency threshold. The only exception to this rule is when it entails wrongful discrimination, in which case states should only follow this rule if victims of discrimination benefit and consent to its implementation. ABSTRACT Introduction States are sometimes permitted to grant asylum to some refugees and not others. Section 2 defends this claim. Section 3 considers how states ought to select the refugees to which they grant asylum. I reject the claim that states ought to select those fleeing persecution before others. For example, it would be wrong for states to always select those fleeing ethnic-based persecution before selecting those fleeing famines, wars, and pandemics. I instead claim that states ought to select refugees in a manner that helps the worst-off refugees and in a manner that maximises the number of refugees obtaining a sufficiently decent life. When helping the worst off conflicts with the goal of maximising the number obtaining a decent life, states ought to select refugees consistent with a principle of priority. Under this principle, states would select refugees to maximise aggregate benefits, with benefits holding greater weight the worse-off refugees are. In Section 4, I argue that such selections may entail wrongful discrimination. This is because sometimes discriminating against certain groups – such as those with particular disabilities – can increase the available resources states have in accepting refugees, leading to more refugees accepted and worse-off refugees accepted. Such discrimination is wrong, but may nonetheless be all things considered permissible if (a) victims of this discrimination benefit compared to nondiscriminatory selection methods and (b) victims of this discrimination consent to such discrimination as a result. When Can States Permissibly Select Refugees? I assume that states have a limit in the costs they are morally required to bear in accepting refugees. For example, ‘Billionstan’ might have a $1 billion budget set aside © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy, John Wiley & Sons Ltd, 9600 Garsington Road, Oxford, OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. 2 Mollie Gerver for providing asylum to refugees, and this budget may be all it has a duty to bear in accepting refugees. If Billionstan could not accept all refugees with this $1 billion budget, then it would need to consider how it selected the refugees it accepted within this budget. It does not necessarily have a duty to accept as many refugees as possible within this budget. This is because some methods of selecting refugees may be morally optimal even if they do not maximise the number obtaining asylum. This will be made clearer throughout. Though I assume states have a duty to bear certain costs, I am neutral as to how much and what these costs are. It may be that states have a duty to accept financial costs that are moderate, but not above some threshold.1 Or perhaps states have a duty to accept as high costs as possible until accepting more refugees endangers citizens’ access to basic rights.2 States may even be permitted to avoid certain cultural costs; perhaps if a very large number of refugees arriving would undermine the state’s culture, the state would not have a duty to accept them all, raising the question of who it must accept. Importantly, there may be imaginary costs, in that the majority of policymakers and voters falsely believe they are permitted to invest no more than a given cost or falsely believe that costs are higher than they are. For example, policymakers and voters might only stand for accepting 30,000 refugees because they falsely believe the costs of accepting more than 30,000 are higher than they are, perhaps falsely believing refugees are a drain on the economy. Or they may be aware of the actual costs but falsely think they have no duty to bear them, perhaps because of xenophobia or racism. When this occurs, there will sometimes be a particular agent tasked with selecting the refugees given asylum, and this agent may be unable to persuade the government to accept more than below-obligatory costs in accepting refugees. This agent should at least select refugees in a way that meets the state’s imagined responsibilities, and the same question arises as to which refugees this agent ought to select. For example, UN Refugee Agency (UNHCR) protection officers in unsafe transit countries are often tasked with recommending refugees for resettlement to safe host states and are unable to persuade host states to invest more costs in accepting refugees, consistent with these states’ real duties. Though such UNHCR officers are powerless to increase costs states will accept, they have power in selecting who is ultimately granted asylum within these lower-than-obligatory costs. For both policymakers and UNHCR officers, then, there is a question of how they ought to select the refugees to which they grant asylum. How to Select Refugees There are five potential selection methods. The final is the most justified. Persecution and Violence Some claim that states are only obligated to grant asylum to those fleeing persecution or general violence. Though states must protect others suffering from grave harm, such as those suffering from malnutrition in low-income countries, they can protect such individuals by sending them aid. In contrast, states cannot save those fleeing © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy Sufficiency, Priority, Selecting Refugees 3 persecution and general violence by sending aid alone. Only providing them asylum will be effective, and so those fleeing persecution and violence have a right to asylum which others do not have.3 This rule is not precise: sometimes a state can only save individuals suffering from poverty by granting asylum because their home governments are blocking international aid.4 But if a state can save individuals through aid, it has no duty to provide asylum. There is a problem with the above argument. Even if states have no duty to accept individuals whom they assist abroad, they may still have a duty to accept individuals who they can but do not assist abroad. If states can send aid and don’t – and there is evidence they don’t5 – many living in poverty can only save themselves through asylum. If individuals can only save themselves through asylum, it seems such individuals have equal rights to asylum as those fleeing persecution and general violence.6 Importantly, even if states do send as much aid as they have a duty to send, states may still have duties to provide asylum to others who have not obtained this aid. For example, if a state fulfils its duties by giving £1 million in aid, saving 50,000 individuals abroad, it might still have a duty to accept additional individuals above these 50,000 if this incurs no extra costs. This would incur no extra costs when refugees pay taxes greater in value than the services they receive. If a state incurs no costs in accepting these refugees, it cannot claim it has no duty to provide them asylum based on earlier assistance alone. Of course, the initial costs of integrating refugees may be high, even if these costs are offset through taxes at a later time, and there may be nonfinancial costs, such as the culture of a country suddenly changing. Even if providing refugees asylum does entail short-term or cultural costs, states may still have a duty to accept some refugees who have already arrived at the state’s borders. This is because of the Proximity Claim: when an agent has sacrificed resources to help those far away, she retains a duty to invest some costs in rescuing those close by, assuming the costs fall below a given threshold.7 For example, if Beatrice has already given £20,000 to charity, no longer holding an obligation to give more, she might still have an obligation to bring an injured person to the hospital if coming across him on the side of the road. This seems true even if this will cost her £5 in petrol and conflict with a cultural activity scheduled that day.8 Given her close proximity to the injured man, the costs for her seem sufficiently low to create a duty despite her previous donations. If the Proximity Claim is true, then states already fulfilling their duty to assist via aid could maintain a duty to rescue refugees at the banks of their shores or the terminals of their airports. A question then arises of which refugees states should accept, assuming they needn’t accept all. Given that states could accept many who are particularly badly off, and who can only improve their conditions via asylum, there is no reason to suppose states should prioritise those fleeing violence and persecution compared to others. Selecting the Worst Off If many individuals are particularly badly off and can only improve their conditions by migrating, perhaps refugees are all individuals who are particularly badly off and can only improve their conditions by migrating. If this is true, it seems that states © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy 4 Mollie Gerver selecting between such individuals ought to prioritise those more badly off over those less badly off. There is more than one way someone might be more badly off than someone else. X could be more badly off than Y if X is (a) less able than Y to access certain basic rights or (b) is deprived of important rights which Y is able to access. An individual suffering from untreated AIDS and persecution ought to be selected before someone who is healthy and suffering from the same persecution. Similarly, when two individuals are suffering from the same rights violations, the individual suffering from a more severe form of the rights violation could be selected first, given that this individual is plausibly worse off than the other. For example, an individual unable to leave a small detention cell perhaps has priority in obtaining asylum over an individual unable to leave her village. There are, however, some drawbacks to relying on rights alone. One drawback is that there will likely be disagreement over how we rank basic rights. This is a problem when two individuals are suffering from the same number of basic rights violations. For example, it is not clear if the right to basic freedom is more important than the right to life. Moreover, there are often multiple individuals who will suffer similar rights violations to the same degree. When this occurs, we could establish who is worse off by appealing to predictions of subjective wellbeing, rather than objective rights. More specifically, we might compare refugees in terms of their predicted Quality-Adjusted Life Years (QALYs) without asylum. To see how this is done, it is useful to momentarily put the discussion of refugees aside to discuss health care. Health-care providers often use QALYs in deciding how to distribute health-care funds. They do so by first asking a random sample of individuals with particular conditions to rank their quality of life between 1 to 10 and then using the average to establish quality of life in a given year. For example, if the average quality of life amongst a random sample of individuals with untreated AIDS is reported as 5 out of 10, it is assumed that the quality of life of an individual living with untreated AIDS is only half that of an individual with full health. A year with AIDS is therefore equal to 1(.5) = .5 QALYs.9 Once health-care providers establish the number of QALYs found in the year of an average person with a given condition, they can use this information to allocate resources. Traditionally, state health-care providers have allocated resources to maximise QALYs, but QALYs can also be used to prioritise the worst off: if an individual is expected to have only 20 QALYs left in their lives – perhaps 10 years living with AIDS – they could be treated before someone with an expected 40 QALYs left in their lives. States could use QALYs to measure who is worse off in granting asylum, prioritising refugees who can expect fewer QALYs without asylum before those who can expect more QALYs if they remain without asylum. To apply this approach, states must first consider what impacts refugees’ quality of life, asking a random sample of refugees without asylum what their quality of life is. For example, a random sample of refugees would include single parents and nonsingle parents, and the former might report a lower quality of life than the latter. Importantly, certain attributes might intersect to create fewer QALYs: single parents who are from a given ethnic group might have a lower quality of life compared to nonsingle parents from this ethnic group and compared to single parents from outside this ethnic group. Once the average quality of life © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy Sufficiency, Priority, Selecting Refugees 5 of various cohorts within the sample is established, states and UNHCR could use this information to make predictions about the quality of life of various cohorts (including those outside the representative sample). If it turned out that QALYs were predicted to be lowest for single parents from a given ethnic group, they could be prioritised over nonsingle parents from this ethnic group and over single parents from outside this ethnic group. If the above method were adopted, it would need to account for certain ethical and practical risks. Ethically, it might be offensive for a powerful and privileged aid worker or official to ask a refugee what their quality of life is. Unlike medical practitioners who one day will themselves grow old and face medical complications, aid workers and government officials would be asking refugees to rate a life which aid workers and government officials will never experience themselves. As a result, this might further alienate refugees. One way to mitigate this worry, though perhaps not overcome it completely, is to pay refugees to conduct these surveys themselves. There is also a practical concern that must be addressed: refugees have an incentive to lie, claiming their quality of life is lower than it is, to increase their chances of gaining asylum.10 This practical concern can be mitigated with a sufficiently large sample of individuals, such that each individual in the sample would have an infinitesimal chance of benefiting from exaggerating how badly off they are. For example, a single parent from the Bari tribe in South Sudan is likely to have only a small effect on the final QALY predictions for all single parents from the Bari tribe if she is just one individual in a sample of tens of thousands of subjects. Assuming she has some desire to be honest, the very low benefit of dishonesty could mean her answers are accurate predictions of her quality of life. Moreover, even if individuals were biased, the survey could still be used if the bias was likely evenly distributed. For example, if single parents indicate an average quality of life of 2 out of 10 with bias, and nonsingle parents indicate a quality of life of 3 out of 10 with bias, and the unbiased response is 3 out of 10 for single parents and 4 out of 10 for nonsingle parents, the survey would still distinguish between a cohort that is better off and worse off. If surveys do turn out to be unethical or unhelpful, then QALYs could be limited to health-related outcomes, drawing upon existing measures. For example, if existing medical studies from a given country find that the average quality of life living with untreated AIDS is .5, and the average quality of life with mostly full health is .9, a refugee in this country who will live another five years with untreated AIDS would be prioritised over a refugee predicted to live a similar number of years with mostly full health.11 While limiting QALY measurements to medical variables will not capture all elements impacting quality of life, it would at least capture some elements, to be supplemented with considerations of rights: a refugee suffering from AIDS and a lack of rights to mobility would be prioritised over a refugee with AIDS and mobility and a refugee without AIDS or mobility.12 Maximising the Number Gaining a Sufficient Life While the above selection method captures the value of helping those most badly off, it does not capture the value of maximising the number of individuals no longer badly off. In other words, there is value in ensuring that as many people as possible gain a © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy 6 Mollie Gerver sufficiently decent life in terms of subjective welfare, rights protection, or some other currency.13 If there is such a value, this leads to a dilemma: it is not clear what to do when there is a conflict between selecting the worst-off refugees and maximising the number obtaining asylum. For example, consider a recent case involving 6-year-old Syrian refugee Jamal, living in Jordon and diagnosed with Ewing Sarcoma. He arguably had a right to treatment, but he could not access treatment in the Jordanian refugee camp. He met a New York-based lawyer who helped him apply for resettlement in the United States to obtain life-saving treatment.14 The cost of such treatment was likely around $100,000, money that could have resettled at least two healthy refugees instead.15 Though Jamal was likely one of the most at-risk refugees, resettling him meant fewer individuals could gain a sufficiently decent life. It is not clear whether he or two other refugees ought to have been selected. Importantly, sometimes helping the worst-off refugees alters imagined costs, leading to fewer obtaining a sufficiently decent life. For example, New Zealand’s annual resettlement quota is 750 refugees,16 and refugees selected are rarely the worst off.17 It is likely New Zealand has a duty to accept those worse off, but voters may disagree, thinking accepting worse-off refugees would mean New Zealand would be unable to accept even these 750 refugees. If civil servants selecting refugees selected the worst off, ignoring voters’ beliefs, voters might support fewer than 750 refugees given asylum. If this was the case, it is not clear if civil servants ought to accept the worse off for resettlement or a greater number of better-off refugees. Selecting those most badly off can also come at the expense of maximising the number obtaining a decent life when those unlikely to be saved through resettlement are selected. For example, imagine that Jamal would have a 30% chance of survival by the time he reached the United States. Compare him to a 4-year-old girl living in Benishangul-Gumuz in Ethiopia, where the under-five mortality rate is 169 deaths per 1,000 live births.18 We might think there is value in granting her asylum, and this value is not captured by prioritising Jamal. Though she is not as badly off, both can only obtain a sufficiently decent life through asylum, and Jamal’s chances of gaining such a life are lower. Why We Should Not Use Aggregate Benefits as a Tie-breaker Here is one tempting way of deciding between helping the worst off and maximising the number gaining a sufficiently decent life. States could appeal to aggregate benefits, using either QALYs, rights, or some other relevant metric. For example, imagine two refugees would gain 56 QALYs with asylum between them at a combined cost of $100,000, while Jamal would gain 95 QALYs for the same $100,000. In this case, if we are choosing between Jamal (thus helping the worst off), or the two other refugees (thus maximising the number gaining a sufficiently decent life), it would be better from the perspective of QALY-maximisation to resettle Jamal. Similarly, imagine that Jamal would gain three basic rights for the remainder of his life if given asylum, which is predicted to continue for another 60 years – perhaps he would gain the right to mobility, health care, and sufficient food for 60 more years – while two other refugees would each gain only one right for the remainder of their lives, predicted to be for another 30 years each. Jamal would gain the same number of © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy Sufficiency, Priority, Selecting Refugees 7 years while gaining more rights and so should be prioritised before the two other refugees. A reason it is tempting to adopt the above policy is that aggregate benefits have independent value. We tend to think it is better if there is more total subjective welfare than less, and more rights protected than fewer. For example, it is better to create a world of greater equality with more welfare and rights compared to a world of greater equality with less welfare and rights. Similarly, it is better to help the worse off in a way that ensures more welfare and rights rather than fewer; and better to ensure that all have a sufficiently decent life in a manner that ensures more welfare and rights rather than fewer. If there is value in aggregate welfare and rights, and we are faced with a choice between helping the worst off and increasing the number gaining a sufficiently decent life, it makes sense to appeal to this aggregate benefit. But there is a problem with appealing to aggregate benefits as a tie-breaker. It entails sometimes helping individuals who are much better off at the expense of far more individuals who are much worse off. Imagine there is one child who will live just below the sufficiency threshold without asylum, and who will live an incredibly happy life if given asylum, gaining welfare far above the sufficiency threshold. In contrast, three other refugees suffering from Ewing Sarcoma will die shortly without asylum, and live an insufficiently decent life with asylum, but still much better lives than without asylum. It seems wrong to prioritise the one over the three. The fact that the one refugee will live an incredibly happy life is important, but the extra QALYs she gains above the sufficiency threshold seem less important than the extra QALYs the other three gain up until the sufficiency threshold.19 This seems like a problem even if aggregate benefits come at the expense not of helping more refugees, but of helping a much worse-off refugee. For example, consider the following two South Sudanese refugees who I interviewed in Kenya. Gatbeel managed to obtain employment in Nairobi, enabling him to access shelter and enough food to be relatively healthy: he eats one or two meals a day consisting of mostly rice, sometimes avocado, and less frequently meat, but has insufficient nutrients according to World Health Organisation standards, and so he is just below the sufficiency threshold.20 Compare him to Nyaluok, forced to remain in Dadaab refugee camp with even less food and lacking treatment for a life-threatening respiratory disease.21 Both are seeking transport to Uganda, where Gatbeel will have full rights and Nyaluok will not: she will lack enough nutrients according to World Health Organisation standards, but she will still obtain treatment and more food. In this example, resettling Gatbeel will help him obtain a sufficiently decent life in Uganda, and resettling Nyaluok will not, but Nyaluok is worse off without resettlement. To determine who should be selected, it would seem wrong to resettle Gatbeel. Though he will gain more rights and welfare through asylum than Nyaluok, the rights and welfare Nyaluok gains seem more valuable, given how badly off she is. In such cases, we could abandon aggregate benefits entirely and institute a lottery, with the winner resettled before the loser. But if we think that Nyaluok should be resettled first, because of how badly off she is, then something wrong has occurred if she loses the lottery and is not resettled. If we think that the three refugees with Ewing Sarcoma should be resettled first, because the benefits they gain farther from the sufficiency threshold are worth more, a lottery will not capture the value of these benefits. © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy 8 Mollie Gerver Prioritarianism Here is a selection method that captures the value of these benefits Liam Shields presents a sufficientarian version of prioritarianism. Prioritarianism holds that the better off someone is, the less weight we should give to each benefit obtained. Shield’s sufficientarianism holds that the weight decreases at a greater rate after a sufficiency threshold.22 He calls this a type of ‘nonuniform weighted prioritarianism’. For an example of such nonuniform prioritarianism applied to refugees, imagine one healthy child will: (a) have a sufficiently decent life only if accessing another 40 QALYs, (b) will access another 39 QALYs without asylum, and (c) will access another 70 QALYs with asylum, thus gaining 70 39 = 31 QALYs compared to no asylum. Each of the 31 QALY she gains with asylum becomes increasingly less important, with the first of the 31 QALYs having the most weight, the second of the 31 QALYs having less weight, and so forth. Moreover, the weight decreases to a greater degree after the first QALY, given that the first QALY raises her QALY number from 39 to 40, and is thus the only QALY below the sufficiency threshold of 40. In contrast, imagine that three refugees suffering from Ewin Sarcoma will live another 3 QALYs each without asylum, or 9 QALYs combined, and live 13 QALYs each with asylum, or 39 combined, gaining 39 9 = 30 QALYs combined. All of the 30 QALYs they gain are below the sufficiency threshold, and each has increasingly more weight the farther down from the threshold: the gain from 3 QALYs to 4 QALYs has the most weight, followed by the gain from 4 QALYs to 5, and so forth. Given that all of their QALY gains are below the sufficiency threshold, their QALY gains count for more than the 31 QALY gains for the healthy refugee. The same nonuniform prioritarianism can be applied to the two refugees interviewed in Kenya, Gatbeel and Nyaluok. If Gatbeel gains more benefits than Nyaluok through asylum, but he is already better off than Nyaluok without asylum, his benefits count for less than the benefits Nyaluok gains from asylum, with the benefits above the sufficiency threshold counting for considerably less. By giving greater weight to benefits below the sufficiency threshold, we recognise that benefits matter more for the worst off, but that sometimes a large number of individuals gaining a sufficiently decent life outweighs the value of helping the worst off, so long as the aggregate gains are great enough. If enough refugees will gain a significantly improved life, then the total benefits they gain (even accounting for the lower weight of each benefit) can counteract the benefits the worst off gain (even accounting for the greater weight of each benefit). Some might reject nonuniform prioritarianism in favor of uniform prioritarianism. This view holds that the better off someone is, the less weight we should give to each benefit they obtain, but the rate does not decrease at a faster rate after a sufficiency threshold.23 For example, while the rights Gatbeel gained would be worth less than those that Nyaluok gained, because he is better off, the rights gained above the sufficiency threshold would count for more than under Shield’s nonuniform prioritarianism, because the weight would not decrease at a faster rate above the sufficiency © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy Sufficiency, Priority, Selecting Refugees 9 threshold. Uniform prioritarianism can still explain the intuition that Nyaluok should be resettled first, assuming she is much worse off than Gatbeel, but it might lead to Gatbeel being resettled first if Nyaluok is not much worse off than Gatbeel. While I shall not attempt to determine whether nonuniform or uniform prioritarianism is preferable, one’s broader view of refuge might impact one’s view. The broader view of refuge I have been assuming is that states have special obligations towards outsiders who are particularly badly off. This implies a type of threshold under which an individual is now in the category of “badly off.” Just like we ought to rescue a drowning child if we can at below a given cost, because the child lacks a sufficient right to life without being rescued, states ought to accept those who are starving, persecuted, or suffering other grave harms, because such individuals lack a sufficiently decent life without asylum. If refugees are special because they are below a sufficiency threshold, then assistance under the sufficiency threshold seems especially important, and this explains why gains below the sufficiency threshold increase at a faster rate compared to gains above the sufficiency threshold. While the above is consistent with how refuge is often viewed, some might accept an alternative view: there are no individuals who are refugees, as separate from everyone else, but only people who live on this earth, with some more badly off than others. If states have obligations to invest some costs in helping outsiders, they ought to invest these costs in a manner that balances the value of helping many people with the value of helping the worse off. Uniform prioritarianism balances these two values, and no sufficiency threshold is necessary. While my initial intuition is that Shield’s nonuniform prioritarianism is more compelling in the context of refugees – it does seem that states have weightier duties to save individuals below some sufficiency threshold – I am not strongly committed to this intuition. And it is not very important. What is most important is this: the value of some form of prioritarianism explains why we should sometimes help those who are worse off, but not always. And prioritarianism provides guidance for a range of difficult cases where states or organisations must decide between helping a large number of refugees or a smaller number of worse-off refugees. This general solution of prioritarianism is still abstract: it does not provide precise guidance for all cases, as it does not establish weights of each added benefit. For example, if a refugee has a small amount of food in a refugee camp, while another has slightly more, it is not clear how we weigh the food gained for the first refugee compared to the second. We can merely claim that each increase in food for the first refugee is worth more than each increase for the second. Without precise weights, we must sometimes rely on intuitions; in other words, we must have a sense of which refugees are sufficiently badly off that their gains offset the gains of those who are better off. I believe we do have this sense for cases like that of Gatbeel and Nyaluok. For cases where it is less clear which refugees should be prioritised, we may ultimately need to appeal to lotteries. While a lottery is problematic when it clearly conflicts with intuitions about refugees like Gatbeel and Nyaluok, it needn’t be problematic when we are uncertain who should be prioritised, because we are uncertain how much weight we place on each additional benefit. © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy 10 Mollie Gerver Discrimination Prioritarianism has a drawback: it sometimes involves wrongful discrimination. Indeed, all of the proposed selection methods except for lotteries have this drawback. I explain why and then argue that this discrimination is justified if it increases the odds that victims of discrimination obtain asylum, and they consent as a result. I will then respond to a potential objection. Why Proposed Selection Methods Involve Wrongful Discrimination Discrimination, as I define it here, occurs when an agent treats individuals differently because of a particular characteristic they hold. Not all discrimination is wrong. It needn’t be wrong to treat lower-income individuals differently by providing them assistance before the wealthy. Discrimination is only wrong when certain facts obtain. What these facts are is up for debate. Some claim discrimination is wrong when it excludes members of a disadvantaged group. Others claim discrimination is wrong when it further disadvantages members of an already-disadvantaged group. Still others claim discrimination is wrong when it demeans members of a minority group. For example, imagine a principal segregates Black and White children. Even if the principal lives in a society where Black children are not disadvantaged, and the segregation does not disadvantage them further, she would still be demeaning Black students. This is because segregation expresses racism, given its association with racism throughout history.24 I shall avoid delving into the debate over when discrimination is wrong and simply presume that discrimination is at least wrong when all three of the above characteristics are present. If a school refused to admit children from an already-disadvantaged ethnic minority, it would be engaging in wrongful discrimination, as it would be excluding and disadvantaging an already-disadvantaged group, and demeaning them in a manner similar to the principal above. I shall also presume that discrimination can be wrong without racist or sexist intentions if it entails direct statistical discrimination. Such discrimination occurs when someone intentionally treats members of a group differently because they are statistically more likely to have certain characteristics.25 A landlord engages in such discrimination when refusing to accept tenants of an already-disadvantaged group because they are statistically less likely to pay rent on time. Even if the landlord acts based on economic interests alone, she acts wrongly because she demeans members of this minority, excludes them from housing, and places yet another disadvantage in their way. Helping the worst off, increasing aggregate benefits, and promoting sufficiency and priority can all entail direct statistical discrimination. Here is an example from medical care. Until recently there was evidence that African-Americans benefited less on average from beta-blockers in reducing blood pressure compared to non-African-Americans, though they still benefited to an extent.26 If beta-blockers are more effective at treating non-African-American patients, then more patients might gain a sufficiently decent life if all resources for beta-blockers were diverted towards non-African-Americans, denying African-Americans this treatment. Such a policy might also increase total QALYs and funds for the worst-off patients if the worst off happened to be nonAfrican-American. If enough of the worst off benefited, this could also comply with © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy Sufficiency, Priority, Selecting Refugees 11 prioritarianism: if we weighted benefits amongst the worst off as greater than benefits amongst the better off, the weighted benefits from discrimination could be greater than the weighted benefits for nondiscrimination. This policy would nonetheless be wrong, as it would exclude African-Americans from urgent health care, disadvantaging them further, and demeaning them in the process: it can be demeaning to walk into a clinic and be denied health care because of one’s ethnicity. Public health-care systems don’t engage in direct statistical discrimination (as far as I know), but refugee resettlement systems do. The United Kingdom, Canada, and Jordan discriminate against single men of Middle Eastern descent in determining who is resettled or granted asylum, claiming Middle Eastern men are more of a security risk, and so providing them security clearance is more costly.27 Given the costs, states can potentially grant asylum to more individuals with their budgets if they discriminate against these men, increasing the number of refugees obtaining a sufficiently decent life, and possibly helping more of the worst-off refugees obtain asylum. These states nonetheless engage in wrongful direct statistical discrimination if Middle Eastern men are already disadvantaged compared to men of other backgrounds, and if such selection procedures are demeaning; it can be demeaning to be denied safety because of one’s ethnic background.28 Selecting based on who is worse off, and based on sufficiency or priority, could similarly entail discrimination against certain religious groups. A given UNHCR officer might know that if she only recommends non-Muslim refugees for resettlement to Slovakia, Hungary, or India – countries that avoid accepting Muslim refugees29 – these countries would accept more refugees than if forced to accept Muslim refugees. This could lead to more refugees obtaining a sufficiently decent life through asylum, producing aggregate benefits and potentially maximising benefits for the worst off. Such discrimination would still be wrong. A similar type of direct statistical discrimination is related to individuals with disabilities. Many countries select who is resettled by considering who will best adapt, and those with disabilities are sometimes considered less likely to adapt.30 Such policies may increase QALYs if those with disabilities gain a lower quality of life from asylum due to hardships in host countries that nondisabled individuals would not face.31 Even if greater weight were given for benefits gained for those with disabilities, on the grounds that those with disabilities may be worse off without asylum compared to others, the total weighted benefits they gained might still be lower than if nondisabled refugees were resettled at a greater rate. They would therefore not be selected for asylum, and such discrimination would seem wrong. Importantly, voters may think that discrimination against those with disabilities will increase the state’s overall budget because those without disabilities will pay more in taxes. Whether or not this is true, voters may think the extra budget from accepting disproportionately more nondisabled refugees will create costlier duties to accept more refugees. If these voters do then accept more refugees, and this leads to a major increase in benefits for the worst off, such discrimination may promote sufficiency and priority. It would still seem wrong. If such discrimination is wrong, this still raises a question of whether the wrongness of discrimination is enough to deem it all things considered unjustified.32 For, we might think that sometimes the benefits of helping the worst-off justify the wrongness of discrimination.This is a possibility I shall now consider. © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy 12 Mollie Gerver A Proposition When faced with scenarios described in the last subsection, we ought to adopt this rule of thumb: the more victims of discrimination increase their chances of gaining asylum as a result of discrimination, and the more they consent to discrimination as a result, the stronger the reason to implement the discrimination. For example, imagine: (a) a state discriminates against refugees with a given disability, in that 1% of refugees it accepts for asylum have this disability, while 2% of all refugees have this disability; (b) those without this disability are likely to pay more money in taxes, increasing the state’s overall budget and the costs it has a duty to bear in accepting refugees; and (c) as a result, the state grants asylum to more individuals, leading to an increase in the absolute number of refugees with this disability gaining asylum, giving them a greater chance of obtaining asylum than under a nondiscriminatory procedure. This would be the case if the discriminating state accepted 100,000 refugees, including 1,000 with this disability, and had it not discriminated there would be less money overall, giving the state a duty to accept only 40,000 refugees, 800 with this disability. Though the 800 would be 2% of all those accepted, and thus not discriminatory, there would be fewer with this disability granted asylum in absolute terms compared to under the discriminatory policy. As a result, those with this disability would have a lower overall chance of obtaining asylum compared to under the discriminatory policy. The more refugees with this disability increase their chances of gaining asylum due to discrimination, and the more they consent to discrimination as a result, the more reason to adopt this policy. In some cases, this reason might be decisive; if the number gaining asylum were high enough, and the number consenting were high enough, the discrimination would be permissible. The argument in support of this claim begins with a premise relating to consent. Consent’s function is to give individuals autonomy over their lives by successfully dissolving the duty others owe to them. Doing so gives others permission they otherwise would not have.33 For example, if Mary consents to Svetlana entering her home, Mary dissolves Svetlana’s duty to not enter her home, giving Svetlana permission to enter her home. To give consent that successfully dissolves another’s duty, one needn’t necessarily be consenting voluntarily; it is enough that one has as much autonomy as the agent obtaining consent has a duty to provide. For example, imagine Svetlana needs protection from a gang, and so asks Mary to give her refuge in her home until the gang stops pursuing her. Mary agrees on the condition that Svetlana consent to paying a small percentage of her income towards food – nothing very costly, just a small amount to cover Svetlana’s costs which Mary can’t afford herself. Svetlana is coerced into her decision by the gang, and she may even face coercion from Mary, in that Mary will force her to leave if she doesn’t pay a small amount towards food. Though she is coerced, she is provided as much autonomy as Mary has a duty to provide her: though Mary doesn’t give refuge for free or the option of being free from the gang’s pursuit, these are options which are supererogatory or impossible for Mary to provide. If the remaining options – refuge at a low price or no refuge at all – are all those Mary has a duty to provide, and if © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy Sufficiency, Priority, Selecting Refugees 13 Svetlana chooses the first option, it seems she is treated permissibly. If she is treated permissibly by utilising her autonomy to consent to Mary’s arrangement, she is fulfilling the function of consent, and her consent is valid. When a government selects refugees for asylum, it has a duty towards particular refugees to refrain from discrimination. These refugees can dissolve the government’s duty to not engage in discrimination – they can give the government permission to engage in discrimination – if they have as much autonomy as the government has a duty to ensure. They have such autonomy if the government provides them all options it has a duty to provide. This might be the case if the government provides adequate conditions in refugee camps and invests in obligatory costs for resettling refugees. If the government (a) provides all such options, and (b) in doing so provides refugees as much autonomy over their lives as it has a duty to provide, and (c) refugees wish to use this autonomy to accept a discriminatory policy, it seems the government acts permissibly in respecting their choices. In reality, it is rare for a government to provide refugees the autonomy it has a duty to provide. Governments often refuse to invest all obligatory costs in sending aid abroad and accepting refugees for resettlement. In such cases, the government denies refugees options it ought not deny. For example, if it refuses to invest all obligatory costs in supporting refugees in camps, it denies refugees the option to access sufficient food in camps, and if it refuses to invest all obligatory costs in resettling refugees, it denies refugees the option to be resettled. Such a government acts wrongly in instituting a discriminatory policy; though refugees have greater autonomy in being able to choose the discriminatory policy than not, their autonomy is unfairly constrained by a government, and so they have not truly given consent. It would be comparable to a thief putting a gun to a victim’s head, telling her she can give her money or her life: the thief enhances the victim’s autonomy by giving her the option of giving up her money rather than only the option of death, but he has not obtained consent to take her money, because he impermissibly denies her the option of neither giving money nor dying. Even when the government fails to obtain refugees’ consent to discriminate, it might still be better for the government to institute a discriminatory policy than not, if this is what refugees desire. While the government acts wrongly in denying refugees various options and instituting discrimination, this may be better than denying refugees various options and not providing the option of discrimination which increases their chance of asylum. To see why, return to the thief: it is better for a thief to give a victim the option of giving up her money and living, rather than only the option of keeping her money and dying. He acts impermissibly either way, but giving one extra impermissible option seems better for the victim than not giving options at all. It is unlikely that all refugees who are discriminated against will consent to such discrimination, even if many will. Moreover, there will be great variation in benefits. Sometimes discrimination will significantly increase the number of refugees gaining asylum and other times it won’t. For this reason, we ought to make a relative claim: of those refugees who would be discriminated against under a given policy, the more who consent and increase their chances of obtaining asylum under this policy, the more reason governments or organisations (such as UNHCR) have to institute this policy. © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy 14 Mollie Gerver This raises a question of how, in practice, governments or organisations would obtain consent from these refugees. I lack the room to fully explore this question, but it could entail independent monitors asking refugees in camps, rural settlements, and urban centres if they would support a discriminatory policy. Such monitors would need to tell refugees what the chances of resettlement would be under discriminatory and nondiscriminatory policies. A related method would be to distribute a mobilebased survey providing refugees information on the odds of obtaining asylum with the prediction of chances made transparent. In other words, there would need to be both a reliable method of gauging refugees’ opinions and reliable evidence to disclose to refugees about the ways discrimination impacts their chances of obtaining asylum. For now, such information is lacking, as is a method of obtaining refugees’ consent; but if such mechanisms were ever available, they may be ethical methods of determining whether discriminatory selections ought to be implemented. An Objection In another article in this issue, Kieran Oberman argues that some types of discrimination are always prohibited in virtue of discrimination alone. If a state were to select able-bodied refugees more likely to contribute to the economy, increasing the state’s overall revenue, this would be prohibited even if the extra revenue lead to more refugees obtaining asylum. He provides a compelling analogy to support this claim: imagine a hospital provided emergency care disproportionality to higher earners who would pay more taxes later if treated, contributing to health care overall. This would seem to express disdain for lower earners, claiming they are less deserving of treatment on account of the lower salary they earned. Even if the government insisted this was not its intention, its actions could be interpreted in a way that was different than intended. Members of the general public, who already presume that low earners are less deserving, might presume that low earners are similarly denied health care because they are less deserving. If refugees were denied asylum based on their low likelihood of contributing to the economy, the public might similarly presume that such refugees were less deserving.34 If so, then such discrimination would be wrong. One response to the above argument concerns the description of the hospital. There is good reason to suppose that the hospital is not expressing disdain, at least if diverting all extra funds towards health care for low earners. Were it to divert all funds it gained to low earners, it would be expressing the following message: ‘You are deserving of health care as much as high earners, and so we are disproportionality treating high earners so that you can better access health care’. Versions of such reasoning may exist in the United Kingdom, where patients in public hospitals obtain quicker treatment if they pay for private care within these hospitals, with profits contributing towards the National Health Services overall, in theory helping low-income patients obtain treatment. If the low-income patients consent to this arrangement and the arrangement is implemented solely for their benefit, such a policy would be expressing that low-income individuals have a special right to health care.35 The same holds true with refugees. If some refugees are discriminated against but consent, and the discrimination is implemented solely for their benefit, the discrimination needn’t be disdainful. © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy Sufficiency, Priority, Selecting Refugees 15 Of course, discrimination can still be disdainful if premised on false information. Sometimes voters falsely assume that underrepresenting certain groups increases the state’s overall budget to help refugees. For example, if voters assume that those with a given disability are less productive, but they are not, then voters are expressing offensive attitudes towards those with disabilities. If policymakers or UNHCR resettle disproportionately fewer disabled refugees in response to voters’ attitudes, policymakers and UNHCR may be expressing the same offensive attitude as well, even if not intentionally. While such discrimination may be wrong in one way, it might still be all things considered permissible. It is permissible if the agents selecting refugees only have access to a discriminatory procedure that increases refugees’ odds of gaining asylum, or a nondiscriminatory procedure that does not. These agents would be expressing a certain respect for refugees in letting them decide which procedure is implemented. In contrast, if they insisted on nondiscrimination, they would imply the following message: ‘We will deny you the option of choosing how we select refugees to ensure you are not wronged’. There is a certain paternalism implied in this expression, such that consensual discrimination may be better. The above claim is consistent with policymakers trying to counter biases which lead to voters accepting more refugees when certain groups are underrepresented: it would be better if voters were less biased against those with disabilities, those of different religions, and so forth. It remains the case that, until such biases are undermined, there will be scenarios where voters only agree to accept more refugees if some refugees are underrepresented. When this occurs, those who are underrepresented ought to have a say in whether this discrimination occurs.36 This raises the question of whether nonrefugees who are underrepresented ought to have a say as well. Nonrefugees from a given demographic might be demeaned by a policy discriminating against refugees from this demographic. For example, nonrefugees who are disabled may be insulted by discrimination against disabled refugees, feeling it reinforces antidisability sentiments. Nonrefugees who are Muslim may be insulted by discrimination against Muslim refugees, feeling it reinforces Islamophobia. I lack the space to fully address this question, but it seems the consent of refugees has special weight compared to the consent of nonrefugees. Refugees are those who cannot access a sufficiently decent life if denied asylum. Assuming this is a greater harm than the harms nonrefugees face from discrimination, the consent of refugees may be enough to justify this discrimination.37 Summary This article began with a question: if there is a limit to obligatory costs states must bear in accepting refugees, and states cannot accept all refugees with these costs, it is not clear how they should select the refugees they accept. I argued that refugees should be selected in a manner that helps those who are particularly badly off and in a manner that maximises the number of refugees gaining a sufficiently decent life via asylum. When these two goals are incompatible, refugees should be selected based on aggregate benefits, with greater weight given to benefits farther from the sufficiency threshold. An exception should only be made when this entails wrongful © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy 16 Mollie Gerver discrimination, in which case the discrimination is only justified if it increases the odds that victims of discrimination obtain asylum, and they consent to such discrimination as a result. Mollie Gerver, Department [email protected] of Government, University of Essex, Essex, UK. NOTES 1 Matthew Gibney, The Ethics and Politics of Asylum: Liberal Democracy and the Response to Refugees (Cambridge: Cambridge University Press, 2004) ibid at 84; Kyrie Kowalik, ‘Defining refugees in terms of justice,’ Peace Review 29,1 (2017): 68–75 at 71; David Miller, ‘Immigration: The case for its limits’ in A. I. Cohen & C. H. Wellman (eds) Contemporary Debates in Applied Ethics (Malden: Blackwell, 2005), pp. 193–205. 2 Arash Abizadeh, ‘The special-obligations challenge to more open borders’ in S. Fine & L. Ypi (eds) Migration in Political Theory (Oxford: Oxford University Press, 2016); Joseph Carens, ‘Aliens and citizens: The case for open borders,’ Review of Politics 49,2 (1987): 251–273 at 259; Luara Ferracioli, ‘The appeal and danger of a new refugee convention,’ Social Theory and Practice 40,1 (2014): 123–144 at 132; Michael Dummett, On Immigration and Refugees (London and New York: Routledge, 2001) at 14 and 50-52; Anne Stilz, Territorial Sovereignty (Oxford: Oxford University Press, 2019) at 171. 3 Max Cherem, ‘Refugee rights: Against expanding the definition of ‘refugees’ and unilateral protection elsewhere,’ Journal of Political Philosophy 24,2 (2016): 183–205 at 190; Matthew Lister, ‘Who are refugees?’ Law and Philosophy 32,5 (2013): 645–671 at 660; David Miller, Strangers in Our Midst: The Political Philosophy of Immigration (Cambridge: Harvard University Press, 2016) at 83. 4 Gibney 2004 ibid at 84; Julian F. Muller, ‘The ethics of commercial human smuggling,’ European Journal of Political Theory (forthcoming) at 12; Miller 2016 ibid; For a similar argument to defend state’s general right to exclude, see Andrew Altman and Christopher Heath Wellman, A Liberal Theory of International Justice (Oxford: Oxford University Press, 2009) at 172–181. 5 OECD, Development Aid Rises Again in 2016 but Flows to the Poorest Countries Dip, 11 March 2017, accessed on 22 February 2018 at http://www.oecd.org/dac/development-aid-rises-again-in-2016-but-flowsto-poorest-countries-dip.htm. 6 For a similar argument, see Kieran Oberman, ‘Reality for realists: Why economic migrants should not just “go home and wait for assistance,”’ European Political Science 17,4 (2018): 658–661 at 659–660. 7 F.M. Kamm, ‘Rescue and harm: Discussion of Peter Unger’s Living High and Letting Die,’ Legal Theory 5 (1999): 1–44 at 14–17 and F.M. Kamm, ‘The New problem of distance in morality’ in D. K. Chatterjee (ed) The Ethics of Assistance: Morality and the Distant Needy (Cambridge: Cambridge University Press, 2004) at 70. 8 Peter Unger, Living High and Letting Die: Our Illusion of Innocence (Oxford: Oxford University Press, 1996) at 60–61. 9 Sarah J. Whitehead and Shehzad Ali, ‘Health outcomes in economic evaluation: The QALY and utilities,’ British Medical Bulletin 96 (2010): 5–21 at 8–12; William MacAskill, Doing Good Better: How Effective Altruism can Help You Help Others, Do Work That Matters, and Make Smarter Choices about Giving Back (London: Guardian Books, 2015) at 37–38. 10 This same problem is less pronounced with QALYs intended to maximise utility for medical distribution. For example, if one is suffering from cancer, claiming one has a low quality of life might increase the chances of obtaining treatment but, if the treatment is expensive or nonexistent, then this might also decrease their chances of obtaining life-extending palliative care, because the QALYs they gain would seem fewer than someone reporting a higher quality of life. A special thanks to an anonymous reviewer for raising this point. 11 The studies we rely on would need to be ones conducted in locations where the conditions are sufficiently similar to those which refugees find themselves in. 12 In establishing QALYs and rights, we ought to account not only for QALYs or rights that refugees certainly lack, but QALYs or rights that refugees have a greater probability of lacking. An individual with a © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy Sufficiency, Priority, Selecting Refugees 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 17 higher probability of continuing to face rights violations or some number of QALYs over the course of her life should often be resettled before someone with a lower probability of facing these rights violations or QALYs. Urgency matters as well: all else being equal, if one refugee has a 90% chance of being killed this year because of a short-term crisis and a 10% chance of being killed within the course of her life, she should be given asylum before a refugee with a 10% chance of being killed this year and a 95% chance of being killed within the course of her life. Assuming both are resettled over the next 2 years, the first should be resettled during the first year and the second during the second. This raises the question of how we calculate life-time or shorter-term risks. I lack the space to fully address this, but we can at least rely on existing medical studies concerning risks. This is similar to what Liam Shields, in broader discussions on distributive justice, calls ‘headcount sufficentarianism.’ See Liam Shields, Just Enough: Sufficiency as a demand of justice, Edinburgh: Edinburgh University Press 2016, ch. 2. I have changed the name to protect the identity of Jamal, whose case was described in a number of news sources in the United States. William G. Ward, Robert M. Corey, Sylvia I. Watkins, “Economic Costs of Malignant Bone Tumors,” The Burden of Musculoskeletal Diseases in the United States, United States Bone and Joint Initiative: 2014, accessed on 7 May 2020 at https://www.boneandjointburden.org/2014-report/viiiab14/economiccost-malignant-bone-tumors. New Zealand Immigration, ‘New Zealand Refugee Quota Programme,’ Ministry of Business, Innovation and Employment, accessed on 27 May 2018 at https://www.immigration.govt.nz/about-us/what-we-do/ our-strategies-and-projects/supporting-refugees-and-asylum-seekers/refugee-and-protection-unit/new-zeala nd-refugee-quota-programme. Murdoch Stephens, ‘Refugee resettlement and activism in New Zealand,’ Forced Migration Review 54 (2017): 43–45. Children born to mothers who have a shorter interval between births, and who are not breastfed, have an even lower survival rate. See Solomon Gebretsadik and Emmanuel Gabreyohannes, ‘Determinants of under-five mortality in high mortality regions of Ethiopia: An analysis of the 2011 Ethiopia Demographic and Health Survey Data,’ International Journal of Population Research (2016): 1–7. For a similar objection in the context of distribution more generally, see Shields ibid at ch. 2. Interview with Gatbeel, Nairobi, 12 February 2018. Interview with Nyaluok, Nairobi, 12 February 2018. Shields ibid at Ch. 2. Martin Peterson and Ove Hansson, ‘Equality and priority’ Utilitas 17,3 (2005): 299–309 at 301, Weirlich, ‘Utility tempered with equality,’ Nous 17,3 (1983): at 423–439 at 431–3; Derek Parfit, ‘Another defense of the priority view,’ Utilitas 24,3 (2012): 399–440 at 403. This example is from Deborah Hellman in Deborah Hellman, When Is Discrimination Wrong? (Cambridge, MA: Harvard University Press, 2008) at 7–36. For other theories of discrimination, see: Hugh Collins, ‘Discrimination, equality, and social inclusion,’ Modern Law Review 66,1 (2003): 16–43; Benjamin Eidelson, ‘Treating people as individuals’ in D. Hellman and S. Moreau (eds) Philosophical Foundations of Discrimination Law (Oxford: Oxford University Press, 2013); Kasper Lippert-Rasmussen 2014 ibid at 175; Tarunabh Khaitan, A Theory of Discrimination Law (Oxford: Oxford University Press, 2015); Tarunabh Khaitan, ‘Prelude to a theory of discrimination law, in D. Hellman & S. Moreau (eds), Philosophical Foundations of Discrimination Law (Oxford: Oxford University Press, 2013) at 145; Shlomi Segall, ‘What’s so bad about discrimination?’ Utilitas 24,1 (2012): 82–100. Lippert-Rasmussen 2014 at chapter 1 and p. 81 For a discussion on new beta-blockers, and the broader debate on clinical trials targeting a single race, see Kerry Grens, ‘Race-Based Medicine?’ The Scientist, 19 November 2007, accessed on 25 May 2018 from https://www.the-scientist.com/?articles.view/articleNo/25688/title/Race-based-medicine-/. Lewis Turner, ‘Who will resettle Syrian men?’ Forced Migration Review 54(February 2017): 29–31. There might also be wrongful discrimination based on marriage status and number of children. States often resettle families more than singles, because families have children, and children are more vulnerable. This could be demeaning towards those who have chosen to not have children. Or imagine a state selecting refugees more likely to send remittances back to surviving family members in transit states. This might promote prioritarianism, because the weighted benefits for those remaining in camps could be significant, but it entails discriminating against those whose family members have all been killed. © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy 18 Mollie Gerver 29 Dargan Thompson, ‘Slovakia Agreed to Take in 200 Syrian Refugees, But Only if They’re Christians,’ Relevant Magazine, 19 August 2015, accessed on 5 December 2018 at https://relevantmagazine.com/slice s/slovakia-agreed-take-200-syrian-refugees-only-if-theyre-christians; Daniel McLaughlin, ‘Hungary to Help Christians While Rejecting Muslim Migrants,’ The Irish Times, 27 September 2016, accessed on 7 December 2018 at https://www.irishtimes.com/news/world/europe/hungary-to-help-christians-while-reject ing-muslim-migrants-1.2807543; AP in New Delhi, ‘Indian Citizenship Law Discriminatory to Muslims Passed,’ The Guardian, 11 December 2019, accessed on 20 January 2020 at https://www.theguardian.c om/world/2019/dec/11/india-to-bring-in-law-denying-citizenship-to-muslim-migrants. 30 Lina Anani, ‘Refugees with disabilities: a human rights perspective, Refuge 19,2 (2003): 23–30 at 27–28; Elgar Crock, Laura Smith-Khan, Ron McCallum and Ben Saul, The Legal Protection of Refugees With Disabilities: Forgotten or invisible? (Cheltenham, UK: Elgar Adler Publishing, 2017), Ch. 11. 31 The same could be said about individuals suffering from an illness which lowers their quality of life. 32 It is worth noting that all forms of refugee selection entail wrongful discrimination in one sense: countries tend to place a limit on refugees they accept partly because they prioritise their own citizens’ less-urgent interests over refugees’ more-urgent interests, and this makes badly-off refugees even worse off, perhaps demeaning them in the process. But such discrimination against all refugees raises a distinct sort of question: whether citizens have a right to prioritise their own interests over many refugees’ interests. Resolving this broad question requires a broader discussion about states’ rights to priorities their own citizens’ interests, and citizens’ rights to prioritise themselves. I put this aside, and focus only on this: not only does discrimination sometimes pit refugees’ interests against citizens, but it can pit refugees’ interests in avoiding discrimination with other refugees’ interests in being resettled. If discrimination sometimes helps more worse-off refugees gain asylum, then refusing to institute discrimination means refugees will pay the price for this refusal. It is not clear when they should. 33 Larry Alexander, Heidi Hurd, and Peter Westen, ‘Consent does not require communication: a reply to Dougherty, Law and Philosophy 6,35 (2016): 655–660 at 656–657; Larry Alexander, ‘The moral magic of consent (II),’ Legal Theory 2 (1996): 166–174 at 166; Tom Dougherty, ‘Yes means yes: consent as communication,’ Philosophy and Public Affairs 43,3: 224–253 at 226; Kimberly Kessler Ferzan, ‘Consent, culpability and the law of rape,’ Ohio State Journal of Criminal Law 13,2 (2016): 397–439; Heidi Hurd, ‘The moral magic of consent,’ Legal Theory 2 (1996): 121–146 at 131; Victor Tadros, Wrongs and Crimes (Oxford: Oxford University Press, 2017) at 204–5. 34 Kieran Oberman, ‘Refugee discrimination – the good, the bad, and the pragmatic’ Journal of Applied Philosophy (this issue). 35 For a report questioning whether patients benefit, see Sarah Walpole, ‘NHS Treatment of Private Patients: The impact of NHS finances and NHS patient care,’ Centre for Health and the Public Interest, 12 March 2018, accessed on 15 December 2018 at https://chpi.org.uk/papers/reports/nhs-treatment-of-pri vate-patients-the-impact-on-nhs-finances-and-nhs-patient-care/. 36 Some might object not to discrimination, but to the requirement of consent: if a given group of refugees are more likely to gain asylum as a result of discrimination, they benefit from discrimination, and so the discrimination does not require their consent. I do not think this is true. Refugees who are discriminated against may feel demeaned while sitting in a refugee camp as their neighbours are resettled, and this feeling may persist even if their odds of resettlement are greater than in a nondiscriminatory scheme. The pain of being left behind because of one’s disability, religion, or ethnicity may not be easily offset by the knowledge that one’s odds of being resettled have increased. 37 It is important to note that, when potential victims of discrimination do not consent to or benefit from discrimination, this needn’t mean they should be resettled. They should not be resettled if the reason they are not resettled is unrelated to the discrimination itself. It would be unrelated to the discrimination if they were rejected for asylum because of a lottery. I claimed earlier that a lottery might be justified when we are uncertain which refugees selected would best promote priority, but a lottery might also be justified if selecting one set of refugees would not promote priority – or indeed the other values I described – and selecting the other set would promote these values but be impermissibly discriminatory. If all refugees from both sets have equal claims to asylum, a lottery could fairly determine who is given asylum while avoiding discrimination. © 2020 The Authors. Journal of Applied Philosophy published by John Wiley & Sons Ltd on behalf of Society for Applied Philosophy