Direct and structural injustice against refugees

Abstract

The dominant philosophical approach to understanding the moral duties that states in the Global North have toward the 26 million refugees worldwide is what we can call the Duty of Rescue Approach.¹ According to this approach, states in the Global North (hereafter Northern states) are mere innocent bystanders overlooking the humanitarian crisis of refugee displacement unfold, and these states have moral duties to rescue refugees from this situation, at least if such states are able to do so at little cost to themselves.²

Serena Parekh's recent normative analysis (2017, 2020) has sought to challenge this dominant approach. Parekh highlights certain Northern state policies and practices used in response to refugees while they are displaced and suggests that refugees endure extensive harms as result of such policies and practices, including the harms of containment and encampment, and their being prevented from accessing adequate refuge. These harms, Parekh argues, are an injustice. Thus, for Parekh, certain Northern states, far from being mere innocent bystanders, are responsible for injustice against refugees.

In this article, I fully endorse Parekh's claims that refugees endure certain harms as a result of Northern state practices, and that such harms constitute an injustice against refugees. Yet, I will explore how we ought to understand this injustice. I contest Parekh's claim that the harms refugees endure as a result of Northern state practices are, and ought to be understood as, a structural injustice—an unfortunate, unintended unjust outcome resulting from structural processes (call this Parekh's Structural Injustice Approach). Instead, I contend that these harms are, and ought to be understood as, a direct injustice against refugees—an unjust outcome directly resulting from specific and avoidable policies enacted by relatively unconstrained actors (call this the Direct Injustice Approach). I argue that Parekh's Structural Injustice Approach fails to accurately capture the causal and normative relations between Northern state practices and the harms endured by refugees, and that this approach fails to provide any advancement on, and suffers from same the problems as, the standard Duty of Rescue Approach to which it is ostensibly an alternative. I instead advocate a Direct Injustice Approach to understanding the harms that refugees endure as a result of Northern states practices. If these harms are indeed a direct injustice, then responsible Northern states are certainly not mere innocent bystanders, and are not merely involved in structural processes that have an unintended unjust outcome (as on Parekh's Structural Injustice Approach), but are instead directly committing a grave injustice against innocent refugees and thus have urgent negative duties to refrain from unjustly harming the world's displaced.

Section 1 explains Parekh's arguments in more detail. Section 2 revisits Iris Marion Young's (2010) account of structural injustice (on which Parekh's arguments are based) to establish the necessary conditions of structural injustices. Section 3 casts doubt on whether the harms that refugees face due to Northern state practices can accurately be cast as a structural injustice according to the necessary conditions. Section 4 advances normative arguments against understanding these harms as a structural injustice, since such an understanding will (among other shortfalls) fail to provide any advancement on the Duty of Rescue Approach and will fail to ground (weighty) moral duties to address the injustice against refugees. Section 5 concludes.

Parekh (2020) criticizes the dominant Duty of Rescue Approach to understanding obligations to refugees on which “[Northern states] are often seen only as rescuers unconnected to the harms that refugees face once displaced.” On this “rescue frame” Northern states “have not done anything wrong. They have not caused refugees to come into harm's way, but are merely stepping in to help” (p. 18). This frame fails to capture the reality of “the harms experienced by refugees and the role that [Northern states] have played in this outcome” and “the harms that we [Northern states] have created” (pp. 19, 158). Parekh, across two books, draws attention to two particular harms that refugees endure as a result of Northern state practices: first the containment and encampment of refugees, and second the inability of refugees to access refuge.

In Refugees and the Ethics of Forced Displacement (2017), Parekh focuses on the containment and encampment of refugees. Northern states, through a variety of policies and practices, have sought to contain refugees in regions in the Global South away from Northern territories. In these regions, refugees are left to reside in refugee camps indefinitely, and Northern states financially and politically support housing refugees in such camps as their preferred response toward refugees (as opposed to resettling or granting asylum to large numbers of refugees; Parekh, 2017, pp. 37–9). The harms of such containment and encampment include “a sense of captivity as well as the denial of freedom, autonomy and basic human rights […] for prolonged periods of time” (p. 5). Refugees in camps are passively dependent on international aid, face anxiety-inducing uncertainty over future prospects, and lack the opportunities necessary for an adequately autonomous existence. Refugees endure such conditions for years, decades, and sometimes generations (p. 3).

Parekh further demonstrates how camps affect refugees' rights. “First refugee camps rarely uphold the rights that refugees are entitled to based on the [1951 Refugee Convention].” Second, “because refugees in camps are so vulnerable, basic human rights are routinely violated both by other refugees and by the state and NGOs, and refugees lack the ability to claim their rights or have violations redressed” (p. 31). Parekh draws upon empirical studies which find the full catalogue of human rights violations in certain camps and conclude that the very structure of camps—as enclosed spaces, beyond the rule of law, that deny free movement—entails that enclosing refugees in camps cannot be reconciled with respecting their human rights (Verdirame et al., 2005). Parekh notes that among the most severe and pervasive violations in camps is sexual violence. “Domestic violence, sexual exploitation, and various kinds of sexual torture occur at extremely high rates.” This “is known to occur globally in all camp settings” (2017, p. 34).

In No Refuge (2020), Parekh focuses on the harm of refugees being unable to access refuge. Parekh understands refuge as “the minimum conditions of human dignity,” which consist of an adequate standard of living (including food, water, clothing, adequate housing, and medical care) as well as physical security against threats to basic human rights (pp. 11–3). Parekh notes that the vast majority of refugees (86%), once displaced from their states of origin, reside in regions in the Global South where they effectively face three options: spend prolonged periods of time without adequate autonomy or security in squalid refugee camps; live in destitution without formal assistance and face exploitation and human rights violations in urban areas; or risk their lives on dangerous journeys and endure extensive human rights abuses to reach adequate security and subsistence in Northern states. Each of these three options fails to provide the minimum conditions of human dignity, and thus the vast majority of the world's refugees are unable to access refuge (pp. 105–6).

For Parekh, this inability to access refuge is a harm that results from Northern state practices. Northern states, seeking to control their borders, have adopted a variety of policies and practices that serve to contain refugees away from Northern territories and prevent them accessing asylum. Parekh cites examples including pushbacks against refugees at European borders, child separation policies in the United States which deter refugees from seeking asylum, Australian interception at sea and returning refugees to off-shore detention centers in Nauru, the detention of refugees in camps and centers in Greece, an EU arrangement with Libya which detains refugees in centers on the Libyan coast and contains refugees in regions in North Africa, and the EU-Turkey Deal which blocks migratory routes and prevents refugees from arriving in Europe (2016, pp. 121–40). In addition, instead of providing adequate access to asylum, or resettling large numbers of refugees, Northern states have opted for providing and funding refugee camps in the Global South as their preferred response to refugees, which, as we saw above, entail a sense of captivity and extensive human rights violations (p. 105).

Due to these practices, refugees are precluded from effective access to asylum in Northern states, and are consigned to endure the squalid camps or destitution in urban areas in the Global South, or risk their lives on the now increasingly difficult and dangerous journeys to Northern states. Thus, as a result of Northern state practices, “we have created a situation in which the vast majority of refugees are effectively unable to get refuge in any meaningful sense; that is, they are not able to access the minimum conditions of human dignity” (p. 159).

For Parekh, the two above harms—the containment and encampment of refugees, and refugees' inability to find refuge—represent injustices against refugees. Thus, far from being innocent bystanders or mere potential rescuers, Northern states are responsible for injustices against refugees.

Though Young does not explicitly specify the necessary conditions, we can infer from her analysis that a structural injustice is: an unjust outcome that is (1) distinct from harms and wrongs that result from direct individual interaction; (2) distinct from harms and wrongs that are the direct result of actions and policies of states (or other institutions); and is instead (3) the unintended consequence of (4) structural processes constituted by (5) a multitude of uncoordinated agents and policies (6) adhering to morally acceptable ends.

I will now argue that the two harms—containment and encampment, and the inability to access refuge—which refugees endure as a result of Northern state practices are not structural injustices understood in the above way. My argument is comprised of two interconnected components. The first is a theoretical-conceptual claim: these harms do not fit the conceptual criteria of a structural injustice and are more accurately understood as direct injustices. The second is normative: a Structural Injustice Approach problematically distracts attention away from direct injustices that must be addressed, absolves state actors of wrongdoing, and entails a framework of weak responsibilities to address the injustice, which, in turn, provides no advancement on the standard Duty of Rescue Approach.

To pre-empt the implications of this discussion, if the harms that refugee endure as a result of Northern state practices are a structural injustice, then it is merely an unfortunate, unintended outcome, and no Northern state is blameworthy or acting wrongfully or causing direct harms to refugees, and these states will thereby have only weak, discretionary responsibilities to amend structural processes that cause the injustice. By contrast, if these harms are direct injustices, then certain Northern states have indeed acted wrongfully and are directly causing harms to refugees, and we are able to hold primary state actors and policies to appropriate account for the unjust harms caused, and such actors will have urgent and compulsory negative moral duties to refrain from causing this injustice against refugees. It matters greatly, therefore, whether these harms are indeed direct or structural injustices.

On my analysis, the unjust outcomes of the containment and encampment of refugees, and their prevention from accessing refuge, are not the unintended by-product of structural processes distinct from individual wrongdoing or unjust policies, but are instead the direct result of agents and specific, intentional and (importantly) avoidable policies. For example, consider the details of the 2017 EU arrangement with Libya and the 2016 EU-Turkey Deal.

In the EU arrangement with Libya, refugees are intercepted on the Mediterranean Sea and returned to indefinite detention in Libya (Human Rights Watch, 2018b). This arrangement has two primary effects. First: detention: intercepted refugees are placed in detention centers on the Libyan coast, funded by EU states including the United Kingdom (D. Taylor, 2018). Overcrowding and lack of sanitation in these centers has led to starvation, disease (in particular tuberculosis), and death. Refugees, including children, also face grievous maltreatment: being raped, beaten, abused, starved, and even traded as slaves (BBC News, 2018; Human Rights Watch, 2019). Documented footage depicts the torture of refugees being burned, maimed, and electrocuted.³ The second effect is containment: this arrangement blocks refugees from traveling from Libya to the EU where they could otherwise have claimed asylum and found adequate safety. It thereby closes off the main migratory route from North Africa to Europe and so contains refugees in harmful conditions in regions in North Africa where their basic subsistence and security needs are not met, and where they are subjected to extensive human rights violations.⁴

The 2016 EU-Turkey deal aims to stem refugee flows from Turkey to Greece and itself has two primary effects. The first is encampment: refugees who have crossed the Aegean Sea and arrived on the Greek islands are enclosed into camps without adequate supplies of food, shelter, and medicine and face extensive human rights violations; this has caused “immense suffering for asylum seekers” (Human Rights Watch, 2016). In the Moria camp, “the sewage system is so overwhelmed, that raw sewage has been known to reach the mattresses where children sleep” (International Rescue Committee, 2018).⁵ There is also a significant threat of physical violence, with women and even children being subjected to sexual violence (Human Rights Watch, 2018a). The mental toll caused by encampment is sufficiently significant that “many people have attempted to end their lives due to the extreme distress and emotional pain they experience” (Human Rights Watch, 2018a). The second effect of the deal is containment as refugees are blocked from traveling to Greece from Turkey and so this main migratory route to safety in Europe is closed down. As a result, refugees are contained in regions nearer their countries of origin in Turkey, Jordan, and Lebanon where their security and subsistence needs are not met. In Turkey, refugees live in squalid camps and face destitution in urban areas without adequate human rights protection and 80% of refugees live in severe poverty (Human Rights Watch, 2016; UNHCR, 2016, p. 55). In Jordan, there is a “rapid deterioration in living conditions” and a significant number of refugees live in abject poverty (UNHCR, 2017). In Lebanon, 70% of refugees live under the extreme poverty line and “each day represents a monumental struggle to meet the most basic needs of food, water and healthcare.”⁶

The arrangement with Libya and the EU-Turkey Deal directly result in the containment of refugees, prevent them from accessing adequate refuge, and consign refugees to endure a lack of adequate subsistence and security in camps and urban areas in the Global South. It seems immediately clear that such harmful outcomes cannot be accurately cast as a structural injustice. The containment policies above are precisely that: policies—devised and then deliberately enacted by states and institutions, with the express purpose of containing refugees. In these cases, the EU council, comprised of representatives from member states, devised, and implemented such policies with the intention of containing refugees and preventing them from arriving on European territories.⁷

To add to these examples, the Australian government policy of interception and return of refugees either to Indonesia or to off-shore detention centers in Nauru or Manus Island under “Operation Sovereign Borders,” is an explicit policy (the latest in a successive trend of off-shoring policies) that forcibly and intentionally denies access to refuge, detains refugees in abusive conditions in centers, and contains refugees away from Australian territory in regions in the Global South where they endure a life in camps (BBC News, 2017; “Operation Sovereign Borders (OSB),”). The US policy of detention and fast-track deportation of asylum seekers, as well as child separation practices, ‘Title 42’ expulsions and ‘Migrant Protections’ turn-back protocols, are again policies intended to deter and prevent refugees from accessing refuge in the United States, and contain those refugees in Central and South America. This forms part of a larger trend of successive deterrence policies enacted by the US against refugees.⁸

Considering these examples, it becomes clear that the containment of refugees and their prevention from accessing refuge is not an unintended, unfortunate, by-product of structural processes, but instead the direct harmful and unjust outcome of explicit and intentional policies specifically designed and enacted by Northern state actors.⁹ Therefore, the resulting harms cannot be understood as structural injustices, but are more accurately understood as direct injustices against refugees.

I now present the second component of my argument: the normative case against the Structural Injustice Approach. On a structural injustice model of responsibility, no one agent or policy or set of policies is directly responsible for causing the injustice, no agent has acted wrongfully, and no policy is itself necessarily unjust, and no agent should be blamed for their contribution to the injustice. Instead, the injustice has resulted from structural processes beyond any individual agent's control or policy contribution. Yet, all agents whose morally permissible actions nonetheless contribute in whatever small way to constituting and reconstituting the structural processes that cause the injustice share responsibility for fixing that injustice (Young, 2010, pp. 100–10). Using this model, Parekh (2017) suggests that responsibility for fixing the structural injustice of containment and encampment, for example, is shared among all those states (and citizens of those states) whose legitimate pursuits of state sovereignty and border controls are nonetheless constituting and reconstituting structural processes that serve to contain and encamp refugees for prolonged periods of time (pp. 122–5).

I argue that this Structural Injustice Approach distracts attention away from directly unjust laws and practices that ought to be the focus for reform, absolves primary actors of (appropriate) moral responsibility and accountability for unjust harms caused, results in an unfair distribution of responsibilities, and entails a problematically weak framework of responsibilities to address the harms against refugees.

First, casting the harms that refugees endure as structural entails that it is difficult to arrive at tangible and actionable objectives when attempting to amend structural processes. How should all those individuals, whose non-wrongful, small-scale actions contribute to structural processes, recognize which and how their actions do so, and then proceed to change them? As Christian Neuhäuser (2014) argues, on the structural model there is no principled means of distributing responsibilities and actions among contributing actors such that “it remains unclear who has to do what” (p. 242). More pressingly, how should Northern states that (on the structural model) are not committing any wrongdoing or adopting unjust policies or directly causing the unjust situation change their ways? There may of course be answers to such questions, but they are not obvious, nor obviously likely to bring substantial change.

For example, promising frameworks of duties and responsibilities have been established by scholars in response to the objection that a structural model of responsibility fails to ground actionable imperatives. Elizabeth Khan (2019) argues that individuals are unable to unilaterally address structural injustices, yet they have a precautionary duty to take action to prevent or mitigate their contributions to ongoing and future structural injustices, which can be discharged through creating and maintaining suitable collectives that are able and willing to address the structural injustice (pp. 41–3). However, applied to the context of understanding state practices against refugees as a structural injustice, this proposal would presumably suggest that individual states are unable to unilaterally address the injustice, but ought to form collectives to address it. This proposal does not obviously yield specific and actional public policy proposals nor clear imperatives or objectives for reform. Robin Zheng (2018) argues that we can fulfill our responsibilities through the performance of our social roles in different ways according a role ideal, which (if done in accordance with others) can help bring about structural change (e.g., teachers can perform their roles in ways that diversify their syllabi to correct for systemic underrepresentation of minorities) (pp. 9–11). Applied to the context of understanding state practices against refugees as a structural injustice, presumably Zheng's proposal would entail that Northern state actors perform their roles in ways that would help mitigate or address structural injustices. This sounds promising. However, as Zheng acknowledges, performing an ideal social role is indeterminate in content and open to each agent's individual conception of the good and what an ideal social role would be (p. 17). As a result, this understanding does not specify which role-performances, policies, or actions would be mandated for Northern state actors, nor establish incentives or parameters for state actors to perform those roles in ways different to how they currently (wish to) perform them. Moreover, since on a structural injustice approach, no state is acting wrongfully, nor are their laws or policies necessarily unjust or directly causing the injustice, there is little grounds here to determine which actions to take or that amending laws or policies are the required actions, nor is there strong moral imperative for reform (a point I shall return to below). Thus, the Structural Injustice Approach, even with both Khan and Zheng's proposals, does not appear to ground actionable reform in the context of state practices toward refugees.

If instead we recognize specific policies that contain and encamp refugees, and deny their access to refuge as directly harmful and unjust, then this provides a clear and actionable objective for change: to amend or reject those policies. For example, states could end the practice of intercepting and returning refugees to abusive detention centers in Libya. The forced encampment of refugees on the Greek islands can be prohibited and the encamped refugees resettled. The containment policies with Libya and Turkey can be identified as harmful and unjust and their reform, to provide safe and legal routes to safety, can be called for. Moreover, current policies that contain refugees in regions where their lives, liberty, and human rights are under threat can be prohibited under International Law as morally equivalent to the (currently prohibited) refoulement of refugees to regions where their lives, liberty, and human rights are under threat, as I have recently argued elswhere (Hillier-Smith 2020). These are actionable imperatives that will demonstrably improve the wellbeing of refugees if and when such policies are properly recognized as direct injustices and reformed. Thus, a structural model risks problematically distracting attention away from these imperatives to the detriment of urgent reform. In fact, one could press the point further and suggest that to be inattentive to these urgent reforms through mis-focusing on structural processes and unclear individual structural responsibilities objectionably fails those refugees whose wellbeing and human rights could be more adequately protected if such harmful practices were identified as directly unjust and prohibited.

Defenders of a Structural Injustice Approach may object and suggest that on their approach, one can point to certain policies and practices (such as those above) that are contributing to the structural processes as appropriate subjects for reform. However, in response, it is unclear why, on a structural model, such policies and practices ought to be subject to reform. After all, none of the policies or practices are themselves unjust but are instead adhering to morally acceptable ends; as Parekh (2020) writes: “structural injustice arises from the actions and policies of thousands of individuals acting according to morally acceptable rules and norms,” and such actions may be “morally neutral or even positive” (pp. 163, 169). Furthermore, on the structural model, none of these actions, policies, or practices are the (direct) cause of the injustice, which is instead the result of structural processes beyond the control and scope of specific actions, policies, and practices. It is hard then to see why, on a structural model, there is an(y) imperative to reform these morally just policies and practices that make little if any difference to the injustice.¹⁴

Second, the Structural Injustice Approach absolves primary state actors of moral responsibility and accountability for unjust harms caused to refugees and thereby “lets them off the hook.” For instance, having detailed numerous practices that Northern state actors have introduced to deny refugees access to Northern territories, including the incarceration of refugees in detention centers in Libya, and the subsequent torture, sexual violence and slavery that refugees face (2020, p. 7), Parekh suggests, that for the harm of refugees being locked out of refuge, “we [Northern states] are responsible not because we have done something wrong [or] something that we should feel guilty about” and that on the structural model of responsibility “we should not blame each other or call each other guilty for [the injustice]” (2020, pp. 170, 171).

This above conception of Northern states' responsibility risks positing that (groups of) states adopting policies such as the EU arrangement with Libya, or the EU encampment and containment of refugees, or the Australian policy of interception and return—which contain and deny refuge and result in significant physical and mental suffering and human rights violations to innocents—are not doing anything wrong or directly harming refugees, but are simply permissibly pursuing their own legitimate interests and ought not to be blamed. This conception objectionably fails to hold these primary state actors and unjust policies sufficiently to account, undeservedly absolves such actors of any wrongdoing, and fails to capture the gravity of their moral responsibility for the significant harms and human rights violations caused to innocent refugees.

Further, on the structural model, responsibility is distributed away from these primary actors, and instead dispersed among many thousands of persons (e.g., any and all citizens of Northern states) whose small-scale contributions have been negligible and non-blameworthy. This unfairly burdens non-blameworthy actors with responsibilities to fix the injustice comparable to those of the primary actors that cause the injustice. This undiscerning distribution of responsibility is therefore unjustly disproportionate not only in absolving the blameworthy actors of proportionate responsibility, but also burdening non-blameworthy actors with disproportionate responsibility, resulting in the unfairness of distributive injustice. By contrast, on a direct injustice model of responsibilities, primary actors, and policies responsible for causing the injustice are held accurately and proportionately to account for the unjust harms caused to refugees, and these agents bear the primary duties to address the injustice.

It may be objected that a direct injustice model is backward-looking, whereas a structural model is forward-looking. While on a direct injustice model, the aim is to trace causal and moral responsibility, assign blame and accountability, and “to demand punishment or compensation”; on the forward-looking structural model, the focus is on addressing the unjust outcome itself and “on how to make things more just in the future” (Parekh, 2020, p. 164). Parekh takes it as an advantage of the structural model that the aim is not to single out particular actors or policies for blame, but to more appropriately focus on the unjust outcome and identify ways that we can work collectively to address it (2020, pp. 162, 170). It may also be objected that assigning blame on a direct injustice model will be counter-productive as one simply produces defensiveness, blame-switching, and resentment on the part of the accused (Young, 2010, pp. 114–7).

In response, a direct injustice model does not preclude forward-looking reform nor addressing unjust circumstances. On the contrary, the aim of identifying certain harmful policies and practices as direct injustices and raising awareness of their unjust consequences is necessarily to seek their reform, and to instantiate more just policies and practices in their place. The aim of identifying persons (in particular state actors) as morally responsible for unjust outcomes is to hold them accountable and as subjects of normative criticism and public moral opprobrium, which in turn aims to incentivize conformity to certain norms and disincentivise their transgression in present and future practice.

Furthermore, as Neuhäuser (2014) has highlighted, producing counterproductive defensiveness similarly arises on the structural model. On a structural model, since no actor is primarily responsible or wants to assume (costly) responsibility for the injustice, there is a strong incentive to avoid responsibilities or pass those responsibilities onto others just as there would be with blame and guilt (p. 244). I further add that there is strong reason to doubt that singling out a policy or actor as directly unjust and subject for moral opprobrium is counterproductive. The identification of an unjust policy or act and resultant public moral outrage can exert a powerful influence and change policy. In fact, the policy of separating migrant families at the border in the United States is one such policy that has since been abandoned in part as a result of being identified as morally unjust and the public moral outrage that followed.¹⁵ Therefore, it is not clear that the structural model can claim an advantage in effectiveness and forward-looking reform.

My last and most crucial claim is that the responsibilities on a Structural Injustice Approach are problematically diffuse and weak. Parekh criticizes the dominant Duty of Rescue Approach on the grounds that, on this approach, Northern states are framed as mere innocent potential rescuers who have done nothing wrong (2020, pp. 158–59) and that states then view the humanitarian crisis of global displacement as not their responsibility and view any obligations to help as discretionary (2017, p. 106). Parekh then aims to provide an alternative to the Duty of Rescue Approach by highlighting the injustices against refugees that Northern states are responsible for, and by advancing a new framework of responsibility to fix these injustices to which Northern states are connected. However, it is not clear that Parekh' approach provides much significant advancement. The responsibility to fix a structural injustice does not appear to be different in kind or have greater weight than duties on the conventional Duty of Rescue Approach, as I shall now explain.

On Parekh's (2017) analysis of containment and encampment, this injustice has occurred as a result of structural processes, which entails that Northern states are not responsible for causing this outcome, nor have they done anything wrongful. As part of the responsibility to fix the injustice, Parekh's account posits that states that are connected and well placed to do so ought to help: “[Northern] states have the capacity to help” and “derive benefits from the phenomenon of encampment by avoiding the costs of settling refugees on their territories.” Parekh suggests that “we [Northern states] have remedial responsibility [in part] based on capacity, effectiveness, and cost […] we have the capacity to aid effectively and at a relatively low cost” (p. 124).

Similarly, Parekh's (2020) conception of responsibility for fixing the injustice of refugees' inability to find refuge “stresses that we are responsible not because we have done something wrong, [and] allows discretion on what we can and should do to address this responsibility” (p. 171). It is true that Parekh suggests that Northern states share responsibility for fixing the injustice since “they played an influential role in how the system is set up,” and “because they have benefited” from the system in avoiding “the burden of hosting large numbers of refugees or processing their asylum claims.” However, Parekh specifies that a “crucial reason” Northern states have a responsibility to fix the injustice is because “we [Northern states] have the capacity to change it” (pp. 173–74). Parekh continues: “in many ways, [Northern] states are like the person watching a child drown in a shallow pond—we are able to do a lot without compromising anything of moral worth” (p. 175).

The above conception of responsibilities appear no different in kind from duties to alleviate the plight of refugees (if one can do so at little cost) on the standard Duty of Rescue Approach. Parekh's structural conception similarly casts Northern states as the innocent bystanders and potential rescuers, not doing anything wrong or directly harming refugees, but who have responsibilities to help insofar as they are able to do so at little cost.

Further, it is not clear that the responsibilities on a structural model are sufficiently weighty to motivate reform, or any weightier than the duties on a standard Duty of Rescue Approach. On the structural model it is hard to see why an agent should be especially compelled to fix the injustice, or why their obligations are particularly binding. After all, on the structural model, the agent did nothing wrong, is not blameworthy, they are not causally responsible for any of the harms caused, and all the harms are the outcome of structural processes beyond their control (Parekh, 2017, pp. 124–5). Why then, for example, ought a state be compelled to help fix a structural injustice against refugees they have not caused, or done anything wrong toward and should not be blamed for, beyond the fact that there is a morally problematic situation and they are in a position to help alleviate it? Thus, it is not clear how, having established that Northern states are connected to an injustice against refugees, the resulting responsibilities are particularly strong, or any stronger or less discretionary than a simple duty to rescue refugees in dire circumstances to which one is not in some way connected. Therefore, the responsibilities on Parekh's approach do not seem any stronger or provide any advancement on the standard Duty of Rescue Approach.

In fact, a more serious problem arises. As Martha Nussbaum (2010) identifies, on the structural model of responsibility, which is strictly forward-looking, if an agent fails to fulfill their responsibility in alleviating an injustice to which they are connected, they ought not to be blamed and have not done anything wrong.¹⁶ Therefore, on Parekh's incorporation of this account of responsibility, states that do nothing to fix the structural injustice are not acting wrongfully and should not be blamed for their omission. This entails that on the structural model, the responsibilities are in fact more discretionary and less weighty than duties on the Duty of Rescue Approach, since on that approach, states have fundamental moral duties to help refugees if they can do so at little cost, and are acting morally wrongfully if they fail to perform such duties.¹⁷ Therefore, it seems that, in fact, the standard Duty of Rescue Approach has the advantage over Parekh's approach in giving rise to weightier obligations to alleviate the plight of refugees, which would be wrongful not to perform. Unfortunately, in seeking to provide an alternative to the Duty of Rescue Approach—by showing how Northern states are connected to harms that refugees endure—Parekh's approach in fact provides a framework of responsibility which is less binding and allows states to fail to aid refugees without acting wrongfully.

Instead, I suggest that once we recognize the harms that result from Northern state practices as indeed direct injustices against refugees, we see that such states are certainly not mere innocent bystanders nor potential rescuers but are directly responsible for unjust harms against refugees. As such these states have primary, nondiscretionary, moral duties to amend or abolish specific policies and practices that cause harm. This Direct Injustice Approach leads to actionable and tangible objectives that have substantive impact. The duties are more fairly distributed in being concentrated on primary perpetrators. Such duties are also much stronger: since states are unjustly harming innocent persons, these states have decisive and compulsory negative moral duties to desist. These duties are particularly strong, stronger than a positive duty to help refugees at little cost; and much stronger than a discretionary responsibility to fix a structural injustice to which the states have non-wrongfully contributed.

Yet, it may be objected that the structural model could be amended to ground more stringent responsibilities. For example, Zheng (2021) argues that agents implicated in structural injustices can be held accountable and subjected to “formative moral criticism” which functions as supportive feedback to encourage them to improve and live up to their responsibilities to address relevant injustices (p. 525). Yet, this formative moral criticism does not involve blame or sanction on agents, nor implies that the agent who failed to live up to their responsibilities has acted wrongfully. This framework then does not appear to ground or motivate stringent moral duties on the part of states to address the injustice that would be wrongful not to perform. In another proposal, Khan (2019) argues that while agents cannot be blamed for past contributions to structural injustice, they each bear a precautionary duty to take action to avoid or mitigate future contributions, for which they can be blamed if they do not perform (p. 40). Similarly, Catherine Lu (2017) suggests that while individuals may not be blameworthy for contributions to a structural injustice, they are blameworthy if they fail to act on their responsibilities to address it (discussed in McKeown, 2021). Thus, Khan's and Lu′s proposals would yield a framework of responsibilities, which would be blameworthy (and presumably wrongful) not to perform. Note however that this amended framework would still take us no further than a standard Duty of Rescue Approach, and would not ground (as) stringent duties as on the Direct Injustice Approach I advocate.¹⁸ Therefore, proposals for an amended model of structural responsibilities may still fail to ground strong duties or stronger duties than the standard Duty of Rescue Approach to address the injustice against refugees. By contrast, the Direct Injustice Approach accurately captures the normative and causal relations between state actors' policies and practices and the unjust harms endured by refugees as indeed direct injustices, and is thus able to ground compulsory negative moral duties to reform or refrain from unjust harmful policies and practices against the world's displaced.

In my view, Parekh's analysis has done more to enhance philosophical understanding of how Northern states are connected to serious harms that refugees face while displaced than any recent work in moral and political philosophy. Parekh's insights have established how Northern state practices result in containment, encampment, and the prevention of refuge and that these harms represent an injustice against refugees. However, I have aimed in this article to constructively critique one aspect of Parekh's important analysis and show that this injustice is not, and ought not to be understood as, a structural injustice. Instead, certain Northern states unjustifiably harm innocent refugees looking for safety through practices of containment, encampment, and denying access to refuge. This is a direct injustice and ought to be identified as such. I personally believe this injustice is one of the most serious in which Northern states are directly implicated in the contemporary era, and that the movement toward the abolition of unjust harmful practices used against the world's displaced ought to be considered an urgent moral priority.