As of the end of 2024, an estimated 123.2 million people worldwide were forcibly displaced due to persecution, conflict, violence, human rights violations, or events that severely disrupted public order. Among them are 73.5 million internally displaced persons, 36.8 million refugees, and 8.4 million asylum seekers (UNHCR, 2025). These figures raise pressing ethical questions, and it is no surprise that Peter Singer, one of the leading ethicists today, has addressed questions of our responsibility towards migrants and refugees in several writings spanning 1988 to 2023. Building on these sources, we aim to reconstruct Singer's ethics of immigration, drawing on both his academic writings and publicly endorsed views. The inclusion of the latter is, we believe, justified insofar as it reflects Singer's role as a practicing utilitarian philosopher and is consistent with positions defended in his academic work. At the same time, this reconstruction should not be understood as presupposing a fully unified or systematically developed theory of immigration in Singer's work. Rather, it involves an interpretive consolidation of positions expressed across different contexts, some of which are fragmentary, and is guided by the aim of making the underlying normative structure explicit. Accordingly, the paper remains interpretive in character and does not attribute to Singer a fully unified or systematically developed doctrine.
At the outset, we identify the core ethical principles and normative commitments of Singer's utilitarianism that will be relevant for the subsequent evaluation of the reconstructed Singerian ethics of immigration, on the assumption that any account of migration must be consistent with his broader moral framework. We then turn to the reconstruction of Singer's ethics of immigration. We first examine his rejection of competing views on moral obligations to refugees. Next, we reconstruct Singer's views on migration and refugee policy. Finally, we critically assess those views, considering, first, their consistency with his general ethical theory and, second, their practical implications relative to existing migration policies. At the end, the paper discusses Singer's critique of the definition of a refugee in the 1951 Refugee Convention and considers whether common critiques of effective altruism extend to his ethics of immigration.
Peter Singer belongs to the utilitarian camp and is one of its most prominent representatives. Initially, Singer defended a Harean form of preference utilitarianism, a position that understands well-being as the satisfaction of rational preferences that an individual would endorse under conditions of full information and critical reflection (Singer, 1993b; cf. Schaler, 2009). According to this view, “an action contrary to the preference of any being is, unless this preference is outweighed by contrary preferences, wrong” (Singer, 1993b, p. 94). This approach, which avoids rule-worshiping (cf. Schaler 2009, p. 387ff), underlies his views on the ethics of immigration.
In his later works, Singer made a slight shift in the understanding of well-being towards hedonistic utilitarianism (De Lazari-Radek & Singer 2014; Singer & De Lazari-Radek 2016), a position based on Sidgwick's understanding that “in any given circumstances the objectively right thing to do is what will produce the greatest amount of happiness on the whole“ (Sidgwick, 2011, p. 200) and the maxim of benevolence that claims “that each person is morally obliged to regard the good of anyone else as much as his own good, except when he judges it to be less, when impartially viewed, or less certainly knowable or attainable by him” (Sidgwick, 2011, p. 186). This shift primarily concerns the redefinition of well-being in hedonistic terms, but it does not affect Singer's endorsement of core utilitarian principles or his broader normative framework, which we briefly sketch in order to assess the consistency of his migration-related views with his central moral commitments. The core principle in Singer's, as in any utilitarian moral theory, is The Principle of Beneficence (PoB): One ought to act in ways that maximize the overall satisfaction of preferences or well-being among all individuals affected by the action.
According to PoB, an action is morally right if it maximizes net satisfaction across all affected individuals. This principle is closely tied to: The principle of the equal consideration of all interests (PoEC): One ought to give equal moral consideration to the interests of all affected beings, without arbitrary preference.
In Singer's own words, “the interests of every being affected by an action are to be taken into account and given the same weight as the like interests of any other being” (Singer, 2002, p. 5). Grounded in the Benthamian formula each to count for one, and none for more than one and Sidgwick's claim that “the good of any one person is no more important (…) than the good of any other; unless there are special grounds for believing that more good is likely to occur in the one case than in the other” (Sidgwick, 2011, p. 186), it reflects widely held views on equality and rejection of arbitrary distinctions. The third principle is famously explicated in Singer's seminal 1972 paper Famine, Affluence, and Morality, where it is stated: “If it is in our power to prevent something bad from happening, without thereby sacrificing anything of comparable moral importance, we ought, morally, to do it” (Singer, 2016, pp. 5–6). It can be formalized as: Duty to Rescue Principle (DoRP): One ought to prevent harm to others when it is in their power to do so, provided they do not thereby sacrifice anything of (comparable) moral importance.
The phrase “without sacrificing anything of comparable moral importance” is further explained as meaning “without causing anything else comparably bad to happen, or doing something that is wrong in itself, or failing to promote some moral good, comparable in significance to the bad thing that we can prevent” (Singer, 2016, p. 6; emphasis AL). In other formulation, Singer is more assertive: “if it is in our power to prevent something very bad from happening, without thereby sacrificing anything morally significant, we ought, morally, to do it” (Singer, 2016, p. 6; emphasis AL). The importance of reducing suffering is emphasized in more recent work by De Lazari-Radek & Singer (2017), who explore the practical and philosophical reasons behind utilitarianism's tendency to prioritize alleviating suffering over increasing happiness. On this view, utilitarian reasoning is most forceful when suffering can be substantially reduced at minimal cost to others, without reducing overall happiness.
Those three core principles of Singer's utilitarianism seem to rest upon four normative claims. The first one is what we would call a: Radical impartiality (RI): Relations, time, and space bear no moral significance.
Singer consistently rejects moral partiality, including the moral significance of spatial, temporal, or relational proximity. Led by the idea that our ethical decision-making should be guided by rational judgment rather than spontaneous emotional reactions, Singer moves beyond customary social norms and spontaneous feelings of empathy towards cognitive empathy. (1) In his famous drowning child thought experiment, he argues that if we are morally obligated to save a child drowning in front of us, we are equally obligated to help distant others (Singer, 2016; 2002). As he puts it: “It makes no moral difference whether the person I can help is a neighbour's child ten yards from me or a Bengali whose name I shall never know, ten thousand miles away” (Singer, 2016, p. 7; 2002, p. 157). Simply put, physical distance is morally irrelevant per se. This commitment to RI also extends across time: “the welfare of future generations should count equally with our own” (Schaler, 2009, p. 323).
RI is central to Singer's ethical framework: it informs PoB (moral rightness depends on maximizing beneficence impartially across all those affected), grounds the PoEC (requiring all interests to be considered equally), and supports DoRP (obliging us to act whenever not sacrificing anything of (comparable) moral importance). This radical tendency towards impartiality is also reflected in the second claim that we recognize as: Omissions Are Also Acts (OAAA): allowing preventable harm is morally equivalent to causing it.
If failing to act results in preventable suffering, it is, according to Singer, morally equivalent to causing harm (Singer, 2016) and no less morally relevant. When discussing moral obligations towards refugees, Singer and Singer (1988) criticized the view that denying entry to refugees outside a country (considered inaction) is morally less significant than forcibly expelling someone who has already arrived (a direct action), arguing that this distinction is misleading, since refusing to admit refugees exposes them to the same risks as sending them back. Therefore, moral responsibility is not contingent on whether harm is caused or merely allowed. Upholding that distinction would conflict with DoRP.
Further, according to Singer, an individual's moral responsibility remains the same regardless of how many others are also in a position to help. As he explains, there is “no distinction between cases in which I am the only person who could do anything and cases in which I am just one among millions in the same position” (Singer, 2016, p. 8). The claim that responsibility is not divisible is what we recognize as the third core claim that Singer's normative framework rests upon: No Split Responsibility (NSR): An individual's moral responsibility is not reduced simply because others have the same responsibility.
In reality, people often treat moral duty as divisible or as dependent on the actions of others. For instance, one may feel less compelled to act if one believes that others are already contributing, or conversely, feel justified in abstaining if others are not contributing. This reflects the understanding of moral acts as supererogatory acts – commendable but not obligatory, where those who abstain are rarely judged or criticized (Singer, 2016, p. 14). However, Singer argues that this understanding is flawed, as it overlooks the moral relevance of avoidable suffering. Under the DoRP and PoEC, acts typically seen as acts of charity are, in fact, acts of duty, and the obligation remains the same regardless of how many others are in a position to help (Singer, 2016, p. 8ff). Others' actions may influence the course of action (according to PoB), but do not provide a moral free pass simply because they might step in or opt out. This brings us to what we recognized as the fourth, and the final Singerian normative claim: No Charity Excuse (NCE): acts that are required by principles PoB, PoEC, and DoRP are acts of duty.
Consequently, failing to act when one can prevent harm without sacrificing anything of comparable moral significance (DoRP) is not just a failure of virtue; it is morally wrong. If these principles and normative claims form the backbone of Singer's ethics, then his account of immigration should be consistent with his broader theoretical commitments. Accordingly, after reconstructing his position, this analysis will evaluate his ethics of immigration in light of that framework. Before turning to the reconstruction, it is also important to note that Singer has been a central figure in the development of effective altruism (EA), an approach grounded in rational, evidence-based, and cost-effective reasoning aimed at maximizing overall well-being. Effective altruism can be understood as an attempt to operationalize utilitarianism in practice (Singer, 2015b; De Lazari-Radek & Singer, 2017; Liu, 2025), thereby offering a basis for assessing migration policy in terms of its feasibility and its effectiveness in promoting the greatest good.
In this section, drawing on his writings from 1988 to 2023, we reconstruct Singer's ethics of immigration. We begin by outlining three argumentative strategies he rejects in this context: the ex-gratia approach to refugee protection, the mutual aid argument, and the all-or-nothing line of reasoning. We then examine his migration policy–related claims concerning the limits of our obligations toward refugees, his characterization of asylum seekers as “queue jumpers”, and his defense of resettlement programs as a realistic and ethically justified alternative to current migration regimes.
In their 1988 paper on the ethics of refugee policy, Peter and Renata Singer criticize the ex-gratia approach to refugee protection. This approach treats refugee admission as a discretionary act of generosity rather than a moral duty, framing protection as an exceptional humanitarian gesture (Singer & Singer, 1988, p. 116). On this view, states have no obligation to admit refugees and incur no moral blame if they refuse to do so. Singer and Singer argue that this position is morally flawed because it conflicts with the principle of the equal moral worth of all human beings, a principle that liberal states claim to uphold (Singer & Singer, 1988, p. 116).
Moreover, the ex-gratia approach is incompatible with the principle of equal consideration of interests (PoEC). As they argue, “immigration policy in general, and refugee intake in particular, should be based on the interests of all those affected, either directly or indirectly, whether as an immediate result of the policy, or in the long run” (Singer & Singer, 1988, p. 121). Drawing on Henry Sidgwick, this view holds that the well-being of all affected individuals (present and future) constitutes the ultimate standard of right and wrong (Singer & Singer, 1988, p. 122). It therefore leaves no moral basis for privileging residents' preferences over those of others affected by a given policy or action. As Carens notes, refugee admission is thus, for Singer and Singer, a matter of moral duty rather than discretionary charity (Carens, 1992, p. 36ff).
This critique of the ex-gratia approach is consistent with Singer's broader ethical framework, most notably PoEC, DoRP, RI, and NCE. For the purposes of the subsequent analysis, his core claims about the ethics of immigration can be formalized as follows: No charity excuse in migration (NCEm): Refugee protection should not be treated as a discretionary act of charity, left to state generosity. Principle of equal consideration in migration (PoECm): Migration policies should give equal consideration to the interests of all affected individuals.
The ex-gratia approach is often supported by the idea of mutual aid, which holds that assistance is a voluntary act grounded in moral connection, solidarity, or shared interests within a community. On this view, a community is bound only by a principle of mutual aid, and its obligations are correspondingly limited: they apply primarily to those who share relevant affinities, while outsiders have little or no claim to protection (Singer & Singer, 1988). In line with their critique of the ex-gratia approach, Singer and Singer also reject this position.
Peter Singer further considers this issue in One World (2002) when addressing the view that citizens may have special obligations toward their compatriots grounded in reciprocity. He not only rejects the mutual aid principle but also emphasizes that, once admitted, refugees can contribute to society in ways comparable to native members, thereby participating in the same reciprocal scheme. Excluding them solely on the basis of non-citizenship is therefore inconsistent with the very logic of mutual aid (Singer, 2002, pp. 169–170). Building on these arguments and drawing on the requirement of radical impartiality (RI), Singer further suggests that national borders lack intrinsic moral significance (cf. Singer, 2002, pp. 169, 176; Singer & Singer, 1988, p. 117). This position, conceptually supported by RI, may be formalized as follows: Principle of moral irrelevance of borders (PoMIB): borders have no intrinsic moral significance and should not determine the weight of our moral obligations.
Finally, Singer also rejects the all-or-nothing line of reasoning. According to this view, a political community has no moral obligation to assist refugees unless it can assist all of them, thereby treating partial assistance as morally insufficient and unnecessary. Singer rejects this reasoning, arguing that it conflates the ideal with the morally required: although it may be impossible to help every refugee, helping some is always morally preferable to helping none (Singer & Singer, 1988, pp. 120–121). The inability to fully resolve a problem does not absolve a community of the responsibility to address it to the extent possible. By rejecting this fallacy, Singer underscores that moral obligations are continuous and scalable, a conclusion consistent with the PoB, DoRP, and OAAA.
The views and positions that emerge from Singer's rejection of these three classical approaches to responsibility toward refugees and migrants are further confirmed in his positive claims about how migration policies should be organized. These claims are examined in the following chapter.
As explained above, PoEC requires that the interests of all affected individuals be considered equally, thereby excluding arbitrary prioritization. In the context of migration, this implies that the interests of citizens and non-citizens must be given equal moral weight. Singer and Singer illustrate this through cases where refugees are excluded on the grounds of potential environmental or social costs, while citizens continue to engage in similar or greater harms. In their view, such reasoning improperly discounts refugees' basic interests in safety and survival in favor of comparatively less urgent lifestyle interests of citizens. This violates PoEC and is therefore, on Singer's account, morally unjustified (Singer & Singer, 1988, p. 126). This illustrates a central commitment in Singer's ethics of immigration: evaluation must begin from PoEC.
Second, PoEC, together with PoB, implies that more urgent interests take priority in conflicts. In most migration contexts, refugees have urgent interests in safety and survival, whereas the costs to residents of admission are typically less severe. This generates a strong prima facie moral obligation for affluent states to admit refugees. However, this obligation is consequence-sensitive, in line with PoB. But, for Singer and Singer, not all the consequences are equally important. They distinguish between definite and speculative consequences (Singer & Singer, 1988, pp. 125–128). Definite consequences are those that can be reasonably expected to follow. For example, if a country accepts refugees, definite consequences will include improved conditions for refugees and increased demand on public resources in receiving states. Speculative consequences, by contrast, include outcomes that are uncertain, such as long-term effects on social cohesion, environmental impacts, or future migration flows. Singer and Singer argue that excluding refugees on the basis of speculative harms is not justified. However, they also allow that speculative risks may become sufficiently probable and serious to count as morally relevant. If, for instance, environmental or social instability reached a level where harm could be reasonably expected, restrictions on admission could be justified (cf. Singer & Singer, 1988, pp. 125–128).
Commitment to this line of reasoning is reaffirmed in Singer's 2015 comments on the refugee crisis, where some previously speculative consequences are treated as more salient and likely, leading him to place greater emphasis on the practical and political risks of increased refugee intake. While Singer continues to regard open borders as a moral ideal, he cautions that unqualified advocacy of them overlooks “our species' lamentable xenophobic tendencies” (Singer, 2015a). Affluent countries, he maintains, have a moral responsibility to accept refugees and should take in significantly more than they currently do (Singer, 2023, p. 286). At the same time, he warns that unrestricted immigration may generate adverse political consequences: it can provoke public backlash, fuel fearmongering, and strengthen right-wing populist movements, potentially resulting in governments that are less supportive of refugee protection, foreign aid, and environmental regulation (Kaczmarek, 2017; Singer, 2018; Singer, 2023). This concern was borne out in the aftermath of the 2015–2016 refugee crisis, which saw the electoral rise of parties such as the Freiheitliche Partei Österreichs in Austria, Lega in Italy, Alternative für Deutschland in Germany, and Prawo i Sprawiedliwość in Poland.
Singer's position, therefore, suggests a moral tension: while strong duties of admission follow from his egalitarian and utilitarian framework, rapid or extensive liberalization of border regimes may undermine the political conditions necessary for sustaining those very commitments over time. The trade-off, in his view, is between expanding refugee admission and preserving the institutional stability required to maintain support for liberal and humanitarian policies. This generates a broader moral dilemma: whether to tighten border controls in order to reduce support for far-right parties, or to risk allowing such parties to gain power and subsequently weaken liberal institutions, including those that underpin refugee protection itself (Singer & Singer, 1988, pp. 127–128; Singer, 2018; Singer, 2023, p. 286). This tension can be formalized in the following way: Limits of Obligation (LO): Our obligation to accept refugees ends where it risks fueling anti-immigrant sentiment and consequently undermining fundamental Western values.
The second normative proposal that we recognize in Singer's writings is also consistent with PoEC, but it shifts the focus from comparisons between citizens and refugees to equal consideration among refugees themselves. On this view, states should grant asylum to those most at risk, regardless of their location (in line with RI), and should give priority to those facing the gravest danger (in line with PoB and DoRP).
It follows that protecting refugees who manage to reach a country's borders, while rejecting others in equally or worse conditions elsewhere, violates PoEC and is therefore morally unjustified (Singer & Singer, 1988, p. 120; Kaczmarek, 2017). This reasoning further implies that turning away asylum seekers can, in certain circumstances, be morally justified. As Singer puts it: “Turning away people who manage to reach one's country is emotionally difficult, even if they are being sent to a safe haven. But we should also have compassion for the millions of people who are waiting in refugee camps. They too need to be given hope” (Singer, 2023, p. 287). This challenges asylum policies that treat territorial arrival as morally decisive.
Singer also conceptualizes asylum seekers in a specific way. In discussing Ethiopian refugees, Singer and Singer ask why those who reach the US border should be allowed to “jump the queue” ahead of others who have waited in refugee camps for resettlement for years (Singer & Singer, 1988, p. 119), thereby emphasizing what they take to be the moral arbitrariness of current asylum systems. At the same time, this reasoning lends itself to the controversial characterization of asylum seekers as “queue jumpers”, a label Singer himself reiterates in later discussions of refugee policy (The i Paper Team, 2016). This view can be formalized as follows: No Special Obligation (NSO): There is no special obligation that states owe to asylum seekers at our borders (queue jumpers) over those waiting in refugee camps.
As an ethical alternative to the current – obviously morally flawed – asylum regime, Singer and Singer (1988, pp. 114–116) discuss three options: voluntary repatriation, local integration in neighboring frontline countries to which refugees first flee, and permanent resettlement elsewhere. They argue that the first two options are largely unrealistic: voluntary repatriation is rarely feasible because the conditions that force people to flee typically persist, while local integration is unrealistic because neighboring states are often economically strained or politically fragile and lack the capacity to absorb large refugee populations. This leaves resettlement as the only viable long-term solution.
In their view, this would involve relocating asylum seekers to safe UNHCR-administered refugee camps, from which structured and safer migration pathways could be organized. These camps and resettlement programs would be supported by wealthy states in coordination with UNHCR, with admission based on vulnerability and urgency. In this way, protection would be allocated according to genuine need, in line with PoEC, while avoidable loss of life and suffering along migration routes would be reduced, in line with DoRP. This position may be formalized as follows: Structured Resettlement Approach (SRA): Relocation of refugees should be done through resettlement programs.
Taken together, LO, NSO, and SRA, reconstructed from Singer's writings on immigration, constitute his positive normative framework for refugee and asylum policy. The following chapter critically assesses the internal consistency and practical implications of this framework.
In this section, we discuss Singer's normative model for refugee policy, asking whether it is consistent with Singer's broader theoretical framework and whether it can serve as a sound alternative to current migration regimes.
Within Singer's framework, rules that typically promote good outcomes, such as assisting refugees, ought to be followed at the intuitive level. At the critical level, however, they may be overridden if their application would foreseeably produce worse consequences. LO operates at this level: it allows that, under certain conditions, admitting refugees may undermine overall well-being by destabilizing the political and institutional conditions necessary for sustaining protection. Political backlash and the rise of anti-immigrant sentiments are thus not intrinsically relevant but become so insofar as they negatively affect aggregate welfare. This is broadly consistent with Singer's rejection of communitarianism; restrictions are not justified by appeals to identity or sovereignty, but by their expected consequences for the stability of liberal institutions and refugee protection (Singer & Singer, 1988). It is also compatible with RI, since it considers long-term effects, including those on future generations.
The problem, however, is that LO sits uneasily with the DoRP, especially in its stronger form. Refugees face immediate and severe harms, such as persecution, violence, and death, whereas the harms invoked by LO are indirect, probabilistic, and mediated through complex political processes. As Singer and Singer themselves concede, the consequences of increased refugee intakes are deeply uncertain (1988, p. 126). And while they are morally relevant, for a sound policy, we would need criteria for assessing their likelihood or magnitude, as well as a clear threshold at which such speculative harms outweigh urgent threats to life. Without such standards, LO risks collapsing into a blanket justification for restrictive policies, grounded in vague predictions about political dynamics, electoral behavior, and institutional resilience. Moreover, it would be in tension with PoEC, NCE, and OAAA.
More fundamentally, LO risks granting xenophobic attitudes indirect normative force. If obligations recede in response to anticipated backlash, the threat of intolerance becomes a reason to comply. In practice, this invites strategic misuse. As debates during the so-called refugee crisis have shown, claims about threats to Western values or political stability are highly malleable and politically exploitable. Rather than enabling careful calibration, LO may incentivize governments to exaggerate risks in order to justify inaction. The likely outcome, therefore, is not a principled calibration of the limits of obligation, but their erosion, accompanied by the normalization of exclusionary practices and a broader rightward shift in political discourse (cf. Muis & Immerzeel, 2017). In fact, this dynamic is already visible within the European Union, where some governments have resisted fulfilling their obligations toward refugees. In response, the New Pact on Asylum and Migration introduced a solidarity and responsibility-sharing mechanism that allows states unwilling to admit refugees to contribute in alternative ways, such as through financial or logistical support. While this approach reflects communitarian reasoning (cf. Wellman & Cole, 2011, pp. 122–123), it underscores a crucial point also relevant for utilitarian migration policy proposals: the central constraint is often not capacity, but the willingness of states to admit refugees.
Moreover, the likelihood of negative consequences is not determined solely by numbers, but also by refugees' perceived identity. States may be more willing to accept culturally or ethnically similar groups while resisting others. If such preferences are incorporated into LO, this would conflict directly with the PoEC and NSO, since admission decisions would no longer be based on need but on the preferences of the host population.
It is also important to consider consequences beyond the receiving state. As Singer and Singer (1988, pp. 124–125) note, refusing refugees may undermine international cooperation, weaken norms of shared responsibility, and damage geopolitical credibility. Whether these consequences should weigh more heavily than the risk of domestic political backlash remains unclear. Furthermore, it could be argued that admitting refugees, even in less welcoming environments, may, over time, reduce xenophobia through increased social contact. Finally, states are bound by international legal obligations to protect refugees. Allowing these obligations to be curtailed on the basis of a loosely defined LO risks undermining the post-war human rights framework that LO itself seeks to preserve.
According to RI, moral obligation cannot depend on location. If two individuals face comparable risks of serious harm, the fact that one has reached the territory of a wealthy state cannot, by itself, justify assigning greater moral weight to their interests. To do so would reintroduce the morally arbitrary distinction that Singer explicitly rejects. PoEC also supports NSO. According to PoEC, the interests of all affected individuals must be treated equally, and moral urgency should depend solely on the severity of harm and the consequences of action. As migration regimes that prioritize arrivals at a country's borders for protection effectively privilege those who, due to greater resources, physical ability, or access to smuggling networks, manage to reach wealthier states, it goes against PoEC; effectively creating a system that benefits the more resourceful rather than the most vulnerable, distorting the moral prioritization of genuine need and leaving those most in danger at a disadvantage. Finally, NSO can also be defended on the grounds of PoB: migration regimes that reward irregular arrivals incentivize dangerous migration routes, exposing even the most vulnerable individuals to harm, exploitation, and even death. From this perspective, denying special priority to asylum seekers at our borders could be justified as a strategy aimed at minimizing overall suffering in the long run. Taken together, those points provide two clear, even though negative, reasons in support of NSO: privileging arrivals at borders would (a) undermine more equitable forms of aid distribution; and (b) further incentivize dangerous (often lethal) migration routes.
But do these two reasons really hold up? Does prioritizing arrivals undermine more equitable forms of aid distribution? If we were dealing with a fixed number of available places, and those places had to be allocated either to asylum seekers at the border or to refugees stranded in war-torn countries, then it would indeed be reasonable to argue that we must impartially assess who has the more urgent needs and prioritize accordingly. This would be justified not only by the principles discussed above, but also by DoRP, which requires that we not sacrifice anything of (comparable) moral importance. If we were choosing between admitting asylum seekers at the border and internally displaced persons or refugees in distant camps with more pressing needs, prioritizing the former would indeed be morally unjustified. However, this dilemma is only theoretical. In reality, most states do not have a strictly fixed number of resettlement places, and even where quotas exist, they are often treated as “ceilings, not goals” (Singer & Singer, 1988, p. 116). It is therefore difficult to sustain the claim that asylum seekers are stealing spots or that admitting them necessarily involves a comparable moral sacrifice in the sense required by DoRP.
This raises a further question: if there is no reliable expectation that states will admit refugees from distant countries with more pressing needs instead of asylum seekers at the border, are they not morally required to admit those at their borders? According to DoRP, we are obligated to prevent harm when doing so does not require sacrificing anything of (comparable) moral importance. Moreover, if a state fails to help the needy at its borders, it would not only violate the DoRP but also the PoB, as inaction itself constitutes a form of causing harm that could have been prevented (OAAA). (2) Finally, addressing the needs of asylum seekers already at the border simplifies the protection process, whereas assisting refugees in distant camps typically requires more complex, resource-intensive procedures, with less certainty that aid will reach those most in need (cf. Carens, 1992, p. 39).
Still, one might argue in favor of NSO, as admitting asylum seekers who arrive at our borders would certainly encourage others to undertake the same dangerous journeys. While perilous migration routes should not be the primary channel of protection, in practice, they are often the only available option. They should therefore not be criminalized but treated as a de facto means of accessing protection (Refugee Council of Australia, 2021), unless the entire migration regime were restructured along the lines of Singer's SRA proposal to provide protection without exposing individuals to such risks.
In the case of such a restructuring, we should be aware of what it might imply. Under the current regime, recognition as a refugee under international law typically entails a right to asylum and limits states' discretion (cf. Lippert-Rasmussen & Vitikainen, 2020). By contrast, a PoEC-based regime could allow states to deny protection even to recognized refugees if other cases are deemed more urgent, effectively creating an abstract refugee status without enforceable rights. Such a shift would significantly weaken existing safeguards, particularly the principle of non-refoulement, “the single most important protection offered to refugees” (Carens, 1992, p. 42) and would effectively reverse key gains of the 1951 Refugee Convention. Refugee status would no longer function as a rights-generating legal category but would be reduced to a merely moral designation. This concern is especially salient given that many signatories to the Convention are already seeking ways to circumvent their obligations (Singer, 1988, pp. 115ff). Dismantling the current framework would therefore risk creating further avenues for evading responsibility, thereby weakening an already fragile international protection regime.
Finally, we would also like to stress that the queue-jumping metaphor, which portrays asylum seekers at borders unfairly bypassing others waiting their turn, is, under current migration regimes, a deeply problematic rhetoric device. As Young (2016; 2018) argues, it operates as a normative system, an allocative mechanism, and a tool of state control. As a normative system, it gives tangible form to the abstract concept of rights by invoking ideas of fairness and order through equal treatment based on time. However, this mirage of fairness is misleading, as access to real-world queues is often shaped by privilege, resources, and entrenched social inequalities. As an allocative mechanism, Young argues, it suggests an orderly and efficient distribution of resources, operating as a legal rule or practice. Yet, this framing oversimplifies the complex legal, political, and economic systems governing access to rights. And, as an instrument of state control, it legitimizes bureaucratic procedures, encouraging compliance and social cohesion while masking deeper structures of exclusion and inequality. In doing so, it depoliticizes human rights and risks undermining their effective realization (Young, 2016; 2018). In the context of migration, the queue-jumping metaphor, first and foremost, obscures the fact that, as Mares' (2001) analysis of Australia's border regime illustrates, there is no queue to begin with. Contemporary migration systems, including asylum and resettlement mechanisms, are decentralized and fragmented. There is no global or universal waiting list; instead, resettlement often depends on arbitrary national policies. And if there is no queue, then there is no possibility of jumping it (Ozoliņš, 2004). Actually, access to refugee status and associated rights depends on a person's ability to reach a particular state's territory (cf. Hathaway, 2021). Under such circumstances, this metaphor reinforces narratives that criminalize asylum seekers, delegitimize their claims, and normalize exclusion. It may contribute to weakening public support for refugees and is frequently deployed in political discourse for problematic purposes (cf. Kentish, 2018). Building on Locke's observation that metaphors and figurative language can distort judgment and manipulate emotions, thereby undermining the pursuit of truth (Locke, 1959, p. 146), we therefore suggest that, within a Singerian utilitarian framework, this metaphor should be rejected.
Resettlement, as an alternative to the current migration regime, would involve relocating asylum seekers to safe, UNHCR-administered camps and providing access to more orderly migration pathways. In theory, this approach offers three key benefits. First, it would reduce exposure to harms associated with irregular migration, including smuggling, trafficking, extortion, forced labor, sexual exploitation, drowning, dehydration, and suffocation. The IOM reports over 60,000 deaths in unsafe migration since 2014, with 2024 being the deadliest year, recording at least 8,938 deaths (IOM, 2025). Second, it would improve prospects for refugees in frontline states, which, as Singer notes, often adopt more restrictive policies (such as limiting services, classifying refugees as illegal, or maintaining harsh camp conditions to deter arrivals) when resettlement is unlikely (1988, p. 115). Third, it may incentivize states to address the root causes of forced migration, thereby reducing destabilizing external interventions.
The resettlement proposal aligns well with Singer's broader ethical framework. It is in line with PoB, requires treating all individuals equally in accordance with PoEC, and, consistently with RI, avoids morally arbitrary distinctions based on geographic location. However, insofar as it permits the rejection of individuals who have already reached the borders of safe countries, resettlement may seem in conflict with DoRP. A response to that objection would be that, in an ideal and fully functioning resettlement system, irregular migration routes would be largely eliminated, thereby mitigating this concern. However, such an idealized system is highly unlikely under current political conditions. And it is the only, but crucial limitation of SRA.
Singer and Singer are right in noting that resettlement is the only viable option for those “who have nowhere to go” and that “millions would choose this option if there were countries who would take them” (Singer & Singer, 1988, pp. 115–116; emphasis AL). The problem, however, is that there aren't. Currently, less than 1% of the global refugee population is resettled annually, and only a small number of 1951 Convention signatories participate in UNHCR resettlement schemes (cf. Garnier, 2023). Moreover, since 2013, approximately 85% of refugees have been hosted by developing countries, highlighting persistent structural imbalances in the global protection regime (UNHCR, 2022).
This suggests that resettlement mechanisms are likely to remain marginal and discretionary, and therefore incapable of addressing the full range of protection needs (cf. de Boer & Zieck, 2020; Parusel, 2021). In addition, these mechanisms operate within a broader restrictive framework characterized by the criminalization of irregular arrivals, including pushbacks, detention, border militarization, and the externalization of asylum responsibilities (Goodwin-Gill, 2014; Lutz, 2020). In practice, resettlement is often used to advance migration control rather than to expand protection (Welfens, 2024), allowing states to project humanitarian legitimacy (cf. Lutz & Portmann, 2021) while simultaneously restricting access to asylum. This reflects what Gibney (2004) describes as the moral schizophrenia of liberal states: endorsing human rights while obstructing access to their realization through legal and administrative barriers. Singer and Singer (1988) acknowledge this tension, noting that despite rhetorical commitments to equality, developed states tend to prioritize the interests of their own citizens over those of refugees, treating assistance as charity rather than as a moral or legal obligation. In such circumstances, resettlement remains the morally justified alternative to the current migration regime, but is also insufficient in practice, given its dependence on the very political structures that perpetuate exclusion.
A reconstruction of Singer's ethics of migration would be incomplete without his two-tier critique of the 1951 Convention definition of refugees. According to the Convention, the term “refugee” applies to any person who, owing to a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group, or political opinion, is outside the country of his nationality and is unable or, owing to such fear, is unwilling to avail himself of the protection of that country; or who, not having a nationality and being outside the country of his former habitual residence as a result of such events, is unable or, owing to such fear, is unwilling to return to it. (UN General Assembly, 1951).
The first tier concerns the first part of the definition, namely its list of grounds that qualify an individual for refugee status. According to Singer, this definition is too narrow and fails to capture the diverse political, economic, and environmental causes of displacement (Singer & Singer, 1988, p. 114; Singer, 2018; Singer, 2023, p. 285). As Singer and Singer argue, “To distinguish between someone fleeing from political persecution and someone who flees from a land made uninhabitable by prolonged drought is difficult to justify, although the latter person would not be classified as a refugee under the U.N. definition” (Singer & Singer, 1988, p. 114). Consequently, Singer challenges the distinction between genuine refugees and economic migrants, calling for a broader definition that includes those displaced by economic and environmental factors.
The second tier concerns the requirement that a refugee is a person who “is outside the country of his nationality, and is unable or, owing to such fear, is unwilling to avail himself of the protection of that country” (UN General Assembly, 1951). According to Singer, there are no moral grounds for prioritizing those who manage to reach another country over those who remain in equally or more urgent conditions elsewhere. Moreover, this approach, according to him, contributes to the emergence of a “new, unscrupulous, and often lethal industry of people smuggling” (Singer, 2023, p. 286).
Both tiers are consistent with PoEC and RI, with the second clearly reflecting concerns underlying SRA and NSO. While the critique raises important ethical questions, its implications warrant closer scrutiny.
Regarding the first tier, we agree that certain other categories of displaced persons also merit recognition. However, broadening the criteria does not necessarily translate into better protection, for three reasons. First, even under an expanded definition, Singer's framework would in practice still tend to prioritize cases of the most immediate and severe harm (PoEC), meaning that the practical distribution of protection would likely remain broadly similar. Second, an expanded definition may undermine the legal clarity and enforceability of refugee status, thereby weakening rather than strengthening the protection regime. And third, there is a risk of misjudging states' willingness to comply. As Carens notes, “We produce less good overall (as defined by the principle of equal consideration of interests) if we expect people to act on the basis of this principle than if we expect them to act primarily on the basis of self-interest and design our institutions accordingly” (Carens, 1992, pp. 36–37). In other words, institutional design that assumes widespread compliance with expanded obligations may yield worse outcomes if states currently reluctant to participate in refugee protection do not, in fact, adjust their behavior accordingly – especially if this leads to the weakening of existing legal frameworks. This highlights a broader methodological tension in utilitarian approaches to migration policy: while moral principles may point toward expansion, effective policy design must remain sensitive to the institutional and political conditions that structure state behavior in non-ideal circumstances.
Regarding the second tier, we agree that the interests of internally displaced persons (IDPs) should not be ignored. However, there are reasons to doubt that revising the current prioritization structure would lead to better overall outcomes. First, the effectiveness of extending protection to IDPs depends on states' willingness to engage in resettlement or comparable forms of external humanitarian assistance. As Wellman notes, contemporary refugee protection has developed partly under “real-world political pressure from leaders who worry about the international legal demands entailed by a more expansive definition” (Wellman & Cole, 2011, p. 119). An expanded definition could therefore generate broader resistance to refugee protection, an outcome that would sit uneasily with Singer's own normative aims (cf. discussion on LO). Second, extending protection to IDPs raises complex questions about the scope of permissible intervention within sovereign states (cf. Wellman & Cole, 2011), especially since refugee status presupposes that individuals are unable to rely on their own government for protection. Third, as already argued in the discussion of NSO, addressing the needs of asylum seekers who reach the border is generally more administratively feasible than identifying and assisting displaced individuals abroad.
A more plausible approach, therefore, may be to adapt Singer's framework to complement rather than replace existing protection mechanisms by addressing the specific vulnerabilities of climate-displaced persons and IDPs without destabilizing the current refugee regime. Developing supplementary forms of protection and strengthening targeted humanitarian interventions may be more practically viable, and ultimately more normatively stable, than replacing a system that, despite its limitations, offers comparatively robust and enforceable safeguards.
Since Singer is widely regarded as a pioneer of effective altruism (EA), it is reasonable to ask whether the criticisms directed at EA apply to his ethics of immigration. At first glance, the connection seems straightforward: both are rooted in utilitarian commitments and focus on reducing suffering. However, a closer examination reveals an important distinction: while many strands of EA prioritize cost-effectiveness as the primary decision-making criterion, Singer's ethics of immigration, as reconstructed in the previous sections (LO, SRA, NSO), is structured around the moral urgency of assisting those who are worst off, constrained by PoEC, PoB, and DoRP.
Two dominant lines of criticism have been directed at EA. First, it is argued that it focuses excessively on individual charitable action, thereby neglecting structural injustices and systemic reforms (Kalajtzidis, 2025). Second, EA's emphasis on measurable impact and economic efficiency is said to reduce complex moral questions to cost-benefit calculations, marginalizing concerns of rights, dignity, and relational responsibility (Kalajtzidis, 2025; Luczak, 2025). The question is whether these criticisms transfer to Singer's migration ethics.
The first critique hardly applies as Singer's migration policy proposals are explicitly structural and political, since migration governance cannot be addressed through private charitable action but requires state-level obligations, legal reform, and institutional coordination. Moreover, Singer explicitly rejects a merely charitable (ex gratia) model of assistance and frames refugee protection as a matter of moral duty rather than voluntary benevolence. In this respect, Singer's migration ethics avoids one of the central weaknesses of EA.
The second critique requires more careful consideration. EA is often understood as asking where resources generate the greatest marginal impact. Applied to migration, this could support efforts to assist refugees in lower-income host countries, where the same resources benefit individuals more. Singer himself acknowledges this consideration, noting that supporting refugees in countries such as Jordan or Lebanon is significantly less costly than providing comparable assistance in wealthier states (Singer, 2015a). Cost-effectiveness, therefore, plays a role in his reasoning.
However, it would be misleading to interpret Singer's migration ethics as reducible to cost-minimization. Cost is introduced as one relevant consideration among others, not as an overriding principle. When Singer argues that affluent states should increase support to less affluent countries hosting large refugee populations, he invokes both humanitarian and pragmatic considerations: reducing hazardous migration, improving prospects for return, and addressing unequal burden-sharing. He also notes that such support “makes economic sense”, citing large disparities in per-refugee costs between Jordan and Germany (Singer, 2015a). Cost, in this sense, functions as a supporting justification rather than the foundation of the moral claim. In the logic of SRA in particular, efficiency considerations are embedded within a broader concern for reducing suffering and preventing harm, rather than replacing moral constraints with aggregate maximization.
If interpreted narrowly as a purely monetary efficiency criterion, EA-style reasoning could risk reinforcing global inequalities by shifting responsibility onto poorer states. However, nothing in Singer's migration framework requires such a reductionist interpretation. Cost considerations operate within a broader normative structure grounded in equal consideration of interests (PoEC) and strong duties to prevent serious harm (DoRP). In this respect, Singer's migration ethics is not a straightforward application of EA, but a more complex position. It partially incorporates efficiency reasoning while embedding it within a utilitarian framework constrained by the normative architecture exposed above. As such, it both engages with and partially withstands standard EA critiques, while avoiding their strongest reductive implications.
We cannot afford to wait for some coming glorious day when everyone will live in loving peace and harmony with each other. Human nature is not like that at present, and there is no sign that it will change sufficiently in the foreseeable future (Singer, 1993a, p. 278).
In this paper, we have aimed to reconstruct and critically assess Peter Singer's ethics of immigration. We began by outlining the core structure of his utilitarian theory, centered on three fundamental principles: beneficence, equal consideration of interests, and the duty to rescue. These principles rest on four underlying normative claims: radical impartiality, the moral equivalence of acts and omissions, the rejection of split responsibility, and the denial of a charity-based understanding of moral duty.
Building on this foundation, we identified and systematized Singer's positions on migration. This included his rejection of the ex-gratia approach to refugee protection, mutual aid reasoning, and the all-or-nothing fallacy, as well as three policy-oriented claims: LO, NSO, and SRA. After reconstructing these positions, we critically assessed them by examining, first, whether they are consistent with Singer's broader normative framework, and second, their plausibility as alternatives to current migration regimes. Our analysis shows that Singer's migration claims, while broadly consistent with his ethical framework, also raise several tensions and limitations. For example, LO risks granting indirect normative force to xenophobic attitudes; NSO raises concerns about legal implications and the use of the queue-jumping metaphor; and SRA, while normatively appealing and in line with Singer's principles, seems too idealistic under current global conditions.
We also addressed Singer's critique of the 1951 Refugee Convention definition, arguing that its expansion would not necessarily lead to stronger or more effective protection for refugees and other displaced individuals. Finally, we assessed the relevance of effective altruism critiques for Singer's migration ethics, suggesting that while certain objections apply in limited form, Singer's framework cannot be reduced to a narrow economistic or individualistic version of effective altruism.
Cognitive empathy is the ability to understand what someone else is feeling or experiencing without necessarily sharing those emotions. It involves perspective-taking – intellectually placing oneself in another's situation – and includes imagining or reasoning about what life is like for the other. In contrast, emotional empathy is the capacity to feel what another person feels. In The Most Good You Can Do, Singer explains that effective altruism relies more heavily on cognitive empathy, as it emphasizes understanding suffering on a broad, impersonal scale and making decisions based on reason, evidence, and impact, rather than being swayed by the emotional pull of a single vivid case (Singer, 2015b, pp. 77–78).
As Carens (1992, p. 39) points out, OAAA is too demanding, requiring more knowledge and ability than we realistically possess. Even when we have direct influence over a situation, such as when refugees are at our borders, excluding them does not automatically equate to deportation, but would depend on where they are coming from and what their alternative options are. However, we will leave the discussion on the plausibility of Singer's core principle for another occasion, as it goes beyond the scope of our current focus.