At the end of the previous chapter, we heard Samir, a young man from a North African country who soon had been held for the maximum time in a Swedish deportation prison, reflect on what would happen to him if authorities did not manage to deport him. Previously, people like Samir, whose applications for protection had been rejected but who resisted deportation, used to be allowed access to shelter and a reduced daily allowance from the Swedish Migration Agency until their deportation could be enforced. However, according to a legal amendment adopted in 2016 as part of the Swedish government’s restrictive turn on asylum and migration, instead, as Samir rightly noted, they would be ‘put on the street’. The 2016 amendment to the Act (1994: 137) on the reception of asylum seekers and others (LMA) withdraws access to housing and social benefits for people who refuse to leave ‘voluntarily’. This amendment was what I was discussing with Hasse, the caseworker at the Swedish Migration Agency quoted above, who worked at one of the agency’s return units. The return units were, bureaucratically speaking, part of the asylum reception system. Yet Hasse’s job was not primarily to provide accommodation but to administer the departure of people whose asylum applications had been rejected – a job that he noted had gained increased political recognition since 2016. Hasse explained to me that the political rationale behind the LMA amendment was that cancelling non-deported people’s access to a bare minimum of social welfare and ‘making them freeze’ would eventually wear them down and make them comply with their deportation order. Reflecting on the implications of the law, Hasse weighed its societal and human costs against the ‘bureaucratically complicated’ option of allowing people whose asylum application had been rejected to remain inscribed in the system. The latter option had been frustrating for officials since, in their experience, people did not care about their deportation order and remained anyway. I take his concluding remark, which he expressed in a rather cynical tone, about the political intention to ‘exhaust’ non-deported people as an invitation to explore the instrumentalisation of access to welfare as a means of deportation enforcement.

Alongside detention, the Swedish authorities use formal abandonment (Davies et al., 2017) and derecording (Kalir and van Schendel, 2017) as means of putting pressure on non-deported people. These techniques signal a shift in governing rationales; in detention and during the bureaucratic preparation of the deportation process, non-deported individuals are ‘meticulously inscribed within the bureaucratic machinery of expulsion’ (De Genova, 2017: 255), with its caretaking as well as coercive functions. In contrast, the Swedish exhaustion strategy operates through the withdrawal of access to welfare services and care, and through bureaucratic derecording (Davies et al., 2017; Kalir and van Schendel, 2017). This governing rationale can be described as a form of ‘violent inaction’ (Davies et al., 2017: 1263), where the non-deported are met with indifference by state authorities; they might be present on the territory but treated ‘as if they no longer exist’ (Kalir, 2017: 66). If the deportation camps in Denmark signalled a shift from meticulous control towards a politics of semi-confinement, deprivation, and selective looking away, the Swedish approach to non-deportable people constitutes a step further along the state violence continuum towards the violence of ‘active inaction’. The policy is justified by a welfare chauvinist imaginary wanting to reserve welfare for ‘natives’ (Keskinen et al., 2016), and by neoclassical assumptions that non-deported people are incentivised to remain in Sweden due to ‘economic pull factors’; accordingly, withdrawing their access to welfare services is not only fair but believed to incentivise them to leave (Mayblin, 2019). The policy approach is also embedded in a narrative of ‘humanitarian deportations’ (Borrelli, 2021: 3484): Swedish state authorities have an explicit preference for ‘voluntary’ deportations, which do not require them to resort to physical force. Thus, by withdrawing access to minimum welfare provisions, authorities are able to put pressure on non-deported people and subject them to injurious conditions, without having to touch them.

Enforcing ‘voluntary’ return

On a rainy October morning in 2017, I was waiting in the entrance hall of the office building in one of the Swedish Migration Agency’s return units. The unit was a former open prison, but with no remnants of fences or surveillance systems, located close to one of Sweden’s larger airports. It offered temporary housing to people whose asylum applications had been rejected while authorities were administering their deportations. Unlike the Danish deportation camps, it was not mandatory for non-deported people to live there. Instead, the accommodation was presented as an offer to those who cooperated with authorities in the deportation case, and for those who refused to leave but whom authorities still thought might ‘change their minds’ and cooperate. Staff working in the accommodation units therefore summoned residents regularly to attend so-called return dialogues, where they discussed different deportation scenarios.

I meet up with the manager of the unit, Kristoffer. We get a coffee and sit down in the staff office area together with two return caseworkers, Mariam and Susanne. A whiteboard covers one of the walls, and it is full of tables and flow charts, which are meant to capture the formal steps in the so-called voluntary departure process. Scattered across the tables are colourful Post-its representing different cases the unit is currently dealing with. The ‘aim’ is to move a case smoothly through the tables towards the end goal: deportation. On the other walls are various posters with motivational slogans for staff (I read, ‘Respect!’ ‘Communication!’), and drawings of staff members – Kristoffer explains that the return unit recently had staff meeting days (planeringsdagar), where officials were asked to describe and evaluate their personality traits. They have kept the posters on the walls, to remind themselves that it is a tough job that requires them to work with themselves continuously. We turn towards the whiteboard with the flow chart, and Kristoffer hands me a leaflet so that I can situate their work in the deportation process better (Figure 5.1).

Figure 5.1 Illustration of the ‘return process’ by the Swedish Migration Agency

He explains to me that the left column illustrates the different stages in the voluntary return process: rejection of an asylum application, repeated return dialogues, perhaps an appeal, followed by more return dialogues, ending with ‘departure from Sweden’ (Resa från Sverige). The right column illustrates a ‘ladder of coercive measures’ that authorities may take if a person does not cooperate with authorities in the deportation case; withdrawing LMA, reporting duty (uppsikt), where a person is obliged to register with the migration agency or the police on a weekly basis (Swedish Aliens Act, chapter 10, § 6–8), detention, and finally, handing over the case to the police for enforcement. As a first step, caseworkers at the return units screen the case to make sure no obvious mistakes have been made in the asylum process. Kristoffer says that their staff have usually worked within the Swedish Migration Agency for some years, and they are used to dealing with challenging cases and situations. This is even more important now, he continues, since ‘the competence of asylum caseworkers is … low. They employed lots of new, young people in 2015, and they got no proper training. So, we see ourselves as the last resort and guarantor that asylum seekers will have their rights respected.’ If an asylum case is rejected, the person can either decide to comply with the deportation order, or appeal to the migration court; meanwhile, the return caseworkers start preparing for the return process. ‘We do not force anyone’, Kristoffer emphasises, ‘we work with voluntary return. Or you can say that it’s voluntary but under the force of law.’ Pointing to the first ‘return dialogue’ column, he explains, ‘for those whose applications have been fast-tracked, like people from the Balkans, we try to discourage them from appealing, as the turnover rates are so low anyway, and we don’t want to give them false hopes. It is better for them to leave as soon as possible.’ Mariam, who has been silent so far, now fills in,

Susanna adds, ‘Yes, we need to make sure they know what the deal is. It might sound mean, but we must be clear: you are not allowed to stay here. We sometimes challenge them and ask, what will you do here? What will you live off? Where will you find a job?’ Kristoffer continues,

I ask what ‘incentives’ they can use for those who are not eligible for assisted voluntary return support, and Kristoffer says,

From Kristoffer, Mariam, and Susanna’s overview, we learn about the different measures available to authorities to pressure people to comply with a deportation order without them having to resort to physical force. We also learn that staff conceive these measures as voluntary – or, voluntary under the ‘force of law’. For this purpose, the officials stressed the importance of making non-deported people know and understand their limited prospects of remaining in Sweden. By framing the issue of deportation as a question of knowing and understanding the right information – that is, the information provided by Swedish migration authorities – responsibility for the deportation case was deflected onto the non-deported people. Caseworkers emphasised the importance of voluntarism, and ‘returning at one’s own accord’ (självmant återvändande) was their preferred terminology to describe this part of the deportation process, before the police intervened to use coercive force.

In 2018, 46 per cent of the 18,761 concluded deportation cases were registered as ‘voluntary’, whereas 44 per cent were handed over to the police for enforcement, either because the person had absconded, or because coercive force was deemed necessary in order to enforce the deportation order (Migrationsinfo, 2019). However, the numbers do not reveal if ‘voluntary’ returns have been enforced with the use of detention, withdrawal of welfare, and so on. All officials I spoke to were aware of this contradiction; some of them merely saw ‘voluntariness’ as part of the bureaucratic lingo. Others were concerned with what the euphemism obscured: namely, the fact that migration officials and non-deported people ‘lived in different realities’, as Palle, a junior caseworker at another return unit expressed it; that many still feared going back to the country they had fled from, and that this fear had not been adequately accounted for in deficient asylum procedures (see Asylum Commission, n.d.). Palle reflected,

In Palle’s view, ‘motivation’ was an ill-disguised threat. Kristoffer told me, ‘to say that we work with motivating measures … I think that sounds a bit like “Arbeit macht frei”’. Or, as Hasse noted, presenting non-deported people with the ‘choice’ of deportation with or without coercive force was ‘a bit like telling them to choose between two evils’. Mariam contended, ‘We can talk nicely to them, or we can try to buy them. That’s it – if there’s no readmission agreement, maybe that’s the final little thing that convinces them to leave. That they can rent a room for a year or so, buy a taxi, or an animal. That’s what we are doing, we’re buying them out.’ Hence, if threats did not work, officials could use extortion.

The migration officials I talked to emphasised that at the end of the day, all they could aspire to do was to ‘help’ non-deported people make ‘the decisions that were best for themselves’, as Hasse put it. The point here is not to question the genuineness of their empathy or intentions. But as we were discussing the officials’ motivation strategies, we kept returning to how the discursive distinctions between ‘voluntary’ and ‘forced’ deportations, between ‘assistance’, threat, and extortion, collapsed in practice. As deportation scholars have highlighted, the language of voluntariness and assistance does important legitimation work for the deportation regime, by obscuring the always underlying threat of force and by disfiguring the de facto ‘absence of viable options’ (Andríjasevic and Walters, 2010: 996) into a semblance of ‘deliberate choice’ (see also Gibney, 2013; Koch, 2014; Lecadet, 2018; Webber, 2011). ‘Voluntary’ deportation procedures are also perceived as a way of enhancing the efficiency of deportations at lower costs, while giving deportation an air of humanitarianism (Bendixsen, 2020). While the rhetoric of assistance was prominent among the Swedish officials, they did not deny the coercive underpinnings of their work. In this way, their attitudes provided a realist contrast to the rosy stories that figure in the official ‘soft deportation’ campaigns by governmental institutions and NGOs that have been dissected elsewhere (Kalir, 2017: 56).

As Kristoffer mentioned, there were also cases where neither threat, nor extortion, nor assistance offers worked. Migration officials at the return units were particularly challenged by the cases where the deportation could not be enforced without the active ‘cooperation’ of the non-deported (e.g., due to a lack of readmission agreements, because deportable people refuse to disclose their identity, or because forced return would violate the principle of refoulement; see SOU 2017:84). These were the cases Kristoffer referred to as ‘catch-22’ situations, and according to the migration officials, they made up a substantial share of the deportation cases they dealt with. Handing over these cases to the police was virtually useless, as Kristoffer explained, since coercive force, incarceration, and chartered planes would not make any difference. And indeed, in their 2018 yearbook, the Swedish Police reported that out of the 10,529 deportation cases handed over by the Swedish Migration Agency to the police for enforcement, 4,980 cases were classified as difficult or very difficult to enforce (Polisen, 2019: 100–101). These people, who remained in Sweden either because they considered other options worse or because they could not be deported, were caught in limbo (SOU 2017:84). Their cases kept being brought up in conversations with migration officials, who referred to them as ‘sourdough cases’, and with police officers. Karin, a senior border police officer, was less concerned about these cases, as she was confident in her ability to ‘wear them down until they comply’. Margareta, on the other hand, who worked in the Swedish Migration Agency’s central organisation (in a ‘quality unit’ that has since been shut down) with – according to her – the Sisyphean task of updating standards and manuals for streamlining the deportation process, answered in a cynical tone when I asked her about these cases,

In these situations, she went on to explain, the fact that non-deported people knew that authorities could not forcibly deport them rendered ‘voluntary’ measures and incentives useless. Margareta’s claim that it is the authorities who are left ‘powerless’ and ‘in limbo’ reflects a bureaucratic logic where non-deportability causes knots or fissures in an otherwise smooth flow diagram. However, with the amendment to the LMA, migration authorities had been provided with a novel strategy to resolve this bureaucratic catch-22: if people could not be deported, they could be derecorded.

Destitution as deterrence

The LMA regulates access to social welfare for people seeking protection in Sweden. Since 1994, people who arrive to seek asylum can choose between being housed in accommodation centres operated by the Swedish Migration Agency (anläggningsboende) and run by municipal or private actors, in which case they receive food and accommodation and a daily allowance of SEK 24 per day for single adults. Or, they may choose to stay in private accommodation (eget boende), usually with family and friends, in which case they may receive a daily allowance of SEK 71 per day to cover food, medicines, and other essential expenses – yet as of 2020, this allowance might be withdrawn if they choose to live in a ‘socioeconomically challenged area’.² In 2019, around 56 per cent of people seeking asylum were estimated to live in private accommodation (Prop. 2019/20: 10). Prior to 2016, people whose asylum applications had been rejected continued to receive a reduced daily allowance and could stay in the Swedish Migration Agency’s accommodation until they left Sweden. However, the amendment to the LMA that entered into force in July 2016 withdraws access to the Swedish Migration Agency’s accommodation units and to daily allowance and subsidised medical care for anyone who does not leave the country within the stipulated deadline for voluntary departure, which ranges between two and four weeks (chapter 8 § 21 of the Aliens Act).

This drastic change in policy was justified with logistical and administrative arguments: when an ‘exceptionally’ high number of people arrived in Sweden to seek asylum in 2015, the Swedish Migration Agency’s accommodation units were filled up. Evicting those whose asylum applications had been rejected, it was argued, would make more space for new arrivals (Prop. 2015/16: 146). The LMA amendment was also supposed to serve a deterrence function, by preventing people from ‘becoming stuck’ in reception facilities and delaying deportation procedures, instead incentivising them to leave (SOU 2018:22). Exceptions could be made in cases where the withdrawal of benefits was deemed to be ‘manifestly unreasonable’ (uppenbart oskäligt): for instance, families with children, and people with disabilities or health conditions of a severe yet temporary nature, could retain LMA until they left the country (Migrationsverket SR 13/2016). Exceptions were also made to those whose deportations were suspended for reasons outside their control (SOU 2017:84). However, the preparatory work of the LMA amendment stated that exception clauses should be used restrictively, and for people deported back to Sweden according to the Dublin Regulation, no exceptions would be made. The law thus rendered access to minimum welfare conditional upon certain requirements: either demonstrating vulnerability, as illustrated by the exceptions made for children and people who were severely ill, or ‘performing’ in accordance with authorities’ wishes by cooperating in the deportation process.

When the law was passed, authorities counted around 12,000 open deportation cases, and 4,300 people were expected to be affected by the amendment. Most of them stayed in private accommodation, and would have their daily allowance withdrawn, whereas those who stayed in accommodation provided by the Swedish Migration Agency received an eviction notice (FARR, 2016). In principle, most caseworkers at the Swedish Migration Agency who worked with deportation processes considered exclusion from welfare services to be a logical consequence of a rejected asylum application. Mariam, for instance, argued that there should be no ‘contradictory incentives’ that might ‘encourage’ non-deported people to remain in Sweden. Others thought that it lent ‘clarity and consistency’ to the deportation process. Margareta told me that the amendment had made the Swedish Migration Agency introduce a new category in their online system SKAPA, which permitted authorities to finalise these cases more swiftly, since no more bureaucratic action (such as granting daily allowance or other welfare benefits) could be taken. The category, I was told, was particularly useful for the catch-22 cases that Kristoffer mentioned earlier, where the person could not be deported by force and refused to comply with the deportation order. Referring to these cases, Susanna explained, ‘we simply put them in a locker. Once we withdraw LMA they simply stay there.’ And, she continued, ‘the police don’t want them in their registers, it’s a political thing, and for us it also looks better: we can complete the case file and send it to the archive’. The LMA amendment thus allowed state authorities to strategically use derecording as a last resort, by placing the people whom they could not deport out of bureaucratic sight and out of politicians’ concerns. However, the amendment also presented officials with new dilemmas. It compelled them to make assessments of who was vulnerable or cooperative enough to retain their basic means of subsistence, and Hasse, who was an experienced return caseworker, told me of a particularly difficult case that he had recently encountered:

According to the migration authorities’ legal guidelines, health-related issues only guarantee continued access to accommodation if the condition temporarily impedes deportation enforcement (see SR 13/2016). In the case Hasse describes, the temporary nature of the woman’s illness could not be guaranteed, since it was likely that she would die from cancer. It is noteworthy that the decisions on the nature of her illness and the decision to make her destitute were taken by migration authorities, and not by healthcare professionals. It is unclear whether the decision makers in this case had considered the fact that her destitution might aggravate her health condition. Alongside the thousands of other people affected by the LMA amendment, the woman in Hasse’s story went from being treated as a subject of care and control to being formally abandoned by state authorities.

In their 2018 report on Sweden, the European Commissioner for Human Rights (2018: 7) wrote, ‘[The LMA] amendment had severe humanitarian consequences for a number of individuals, who ended up living in the streets. It also resulted in increasing numbers of people seeking support and help, notably asking for food, and in a reportedly serious deterioration of health situations.’ NGOs working to support young, non-deported people have similarly reported that the LMA amendment has aggravated the social marginalisation of this group, and increased the burden on civil society, which has filled the gap in welfare provisions (Jansson-Keshavarz, Lundberg, and Obenius, 2021). In 2020, during the COVID-19 pandemic, NGOs providing legal advice and essential services to non-deported people reported having to turn away people with symptoms, knowing they had nowhere else to go (Inci et al., 2020). Yet, the amendment is in line with policies adopted across European countries as part of their deterrence approach to non-deported people. By reducing their access to essential welfare to an absolute, humanitarian minimum, the idea behind the approach is to disincentivise non-deported people from staying, and to strategically make them disappear from the registries of bureaucratic authorities. Research from Norway (Johansen, 2013), the Netherlands (Kalir, 2017; Van der Leun and Bouter, 2015), Switzerland (De Coulon, 2015), Ireland (Lentin and Moreo, 2015), and the United Kingdom (Bloch and Schuster, 2005; Mayblin, 2019) has shown that destitution as a mechanism of deterrence has failed to enhance deportation rates. Instead, and as the EU Commissioner also highlighted, it has pushed non-deported people into destitution, and as a result the responsibility for their bare survival has been outsourced to civil society (Rosenberger and Küffner, 2016).

This formal abandonment of non-deported people, who most likely remain in the country under extremely precarious conditions, needs to be understood not as an unintended side effect of the deterrence policy but as a governing technique in its own right. In academic debates, these approaches have variably been described as a ‘politics of exhaustion’ (Ansems de Vries and Guild, 2018: 1), ‘violent inaction’ (Davies et al., 2017: 1263), and an expression of necropolitics (Mbembe, 2003; see also Mayblin et al., 2019), which permit authorities to ‘solve’ the problem of non-deportability by treating people as if they do not exist. It is an example of indirect state violence, which facilitates harm, without ever needing to touch – and without overtly breaching human rights obligations.³ While exhaustion, derecording, and inaction describe how the policies work, their implications are better captured through the concept necropolitics, which underscores the racialised imaginaries facilitating this form of indirect or slow violence. Ultimately, the politics of ‘letting die’ relies on a racial matrix of human hierarchisation where non-deported people are considered disposable, and as such, can rightly be left in a condition where they are merely (or barely) prevented from physically dying (Mbembe, 2003). Moreover, discourses of voluntarism and responsibilisation are important for understanding how these minimum rights approaches are justified, since they construct non-deported people as the cause of ‘their own vulnerability and exploitability’ (Luibhéid, 2013: 2). Like deportation prisons, the politics of abandonment performs and reproduces the boundaries of membership and human worth.

The dilemmas of derecording

If the LMA amendment was supposed to perform ‘clarity and consistency’ in the exclusion of non-deported people from the welfare state, it also amplified tensions internal to the welfare state apparatus in at least two ways. First, it contradicted the principle of bureaucratic inscription fundamental to the work of deportation enforcement officials.⁴ At the Swedish Migration Agency’s return units, caseworkers argued that making non-deported people destitute would simply make them lose contact with authorities. Susanna said, ‘now with the amended LMA, they don’t have any incentives to remain in the system. What should we do, ask them for their address and then be like ah, you live under the bridge over there? That’s just absurd!’ Similarly, Palle told me that he had experienced non-deported people laughing at him when he had first communicated the decision of the Swedish Migration Agency to make them destitute and then, in the next moment, asked for their address (which they had to provide to authorities as to show that they remained ‘available’ to authorities for enforcement). ‘And’, Palle continued, ‘it’s even worse if we accept someone back according to Dublin. Then we say welcome back, we had to take you back here but now we will put you on the street. And one wonders, what the message is here?’ Some return officials shared that they would make exceptions in cases where rejected people had demonstrated their willingness to cooperate with authorities in their deportation process and allow them to remain in the accommodation unit for some extra nights without registering them. Otherwise, they contended, people would lose their trust and any incentive to remain in contact with authorities. As a discretionary measure, the LMA amendment therefore also allowed authorities to expand their repertoire of enforcement strategies.

In the Swedish deportation prison where I conducted fieldwork, detailed in Chapter 3, staff told me that they had realised that the police were releasing people from detention as a way of pressuring people refusing to cooperate with authorities in the deportation process. Yasmin, one of the detention officials, explained that ‘once there was a guy they had to carry out of here, he didn’t want to leave because he had nowhere else to go … it’s like, you call the police to take them here and then you have to call them again to get them out of here!’ I observed a similar scenario on a Saturday afternoon in late September 2017.

We are on the couch in the common room. Livia walks in, bewildered. ‘There is a man waiting at the reception and he says he wants to return to his country of origin. His asylum application has been rejected and his case has been handed over to the police because he absconded, but now he quarrelled with his girlfriend and has no place to stay.’ Livia explains that the man claims that the police referred him to detention, but she is not sure if that can be trusted. ‘And we obviously can’t detain him’, she says,

The incident ended with the man simply being told to leave, as there was nothing they could do for him. Livia suspected that the police would want to place him under registration duty, but since the incident occurred during the weekend, they probably would not make that decision anytime soon. The officials found the incident ironic: they had people trying to escape from the deportation prison, while others were desperately trying to be incarcerated. The story can be read as a way of banalising incarceration – as Khosravi (2009) has highlighted, detention staff used stories such as these to convince themselves and others that detention was not so bad after all. But it also demonstrates how incarceration and destitution are used as complementary measures to govern non-deported people, for whom law enforcement proliferates into a ‘continuum of liminal spaces’, encompassing direct and indirect state violence, coercive control, surveillance, and formal abandonment (Schmid-Scott, 2018). In a way, the man who asked to enter the deportation prison to access food and shelter challenged the derecording logic, as he demanded recognition by authorities and access to basic welfare provisions. But his action also testifies to the precarity of his condition, where a deportation prison is the only place for him to access minimum social assistance.

The second tension or dilemma that the LMA amendment amplified was within the ‘social arm’ of the welfare state, which was now ascribed an even more prominent role in migration enforcement. Minimum rights policies are examples of how social welfare is instrumentalised for the purpose of ‘internal bordering’ (Tervonen et al., 2018: 139), and social services are mobilised as gatekeepers of essential welfare service provisions. As such, the active role of social services in border enforcement is not a novel phenomenon. Defining categories of deservingness and delineating membership are integral parts of social work, and social services continuously partake in defining the boundaries of the welfare state by determining who can access social support and under what conditions (Spencer, 2016; Tervonen et al., 2018). In the context of border enforcement, research from across Europe has demonstrated how social services have increasingly been mobilised as de facto border enforcement agents, who not only determine access to welfare but who are also obliged to check the immigration status of welfare users, and gather and share their personal contact information with police and migration enforcement authorities (Lundberg et al., 2017).⁵ In these contexts, social workers and welfare providers are confronted with conflicting responsibilities: to guarantee access to basic social rights for anyone present on the territory, on the one hand, and to uphold the boundaries of membership in the welfare state, on the other (Björngren and Staaf, 2014). The LMA amendment needs to be seen in the light of the proliferation of internal bordering practices, and as a rather extreme form of, as Nordling and Persdotter (2021: 155) have phrased it, ‘bordering through destitution’.

At the time when the LMA amendment entered into force, people whose asylum cases were rejected and other people living as undocumented in Sweden could turn to the municipal social services and apply for emergency assistance, which was meant to cover food and basic needs (but not accommodation), in accordance with the Social Services Act (2001:453). Decisions on who was eligible for this support and the amount offered were at the discretion of the social services, and varied greatly between municipalities (Nordling, 2017), but in principle, Livia, Palle, and the other migration officials were right in referring non-deported people to social services as a last resort when they lost their LMA. However, in a 2017 ruling, the Supreme Administrative Court (HFD 2017 ref. 33) concluded that social services were not obliged to provide support to people whose asylum applications had been rejected and who were at risk of deportation – including families with children. The court stated that these people should receive assistance from the Swedish Migration Agency, which the LMA amendment effectively precluded. Consequently, non-deported people lost access to any legally guaranteed social support. Social services still retained a certain possibility to grant them emergency assistance to ‘avert dangers to life and health’ (in accordance with chapter 2, § 1–2, SoL), yet at their own discretion (according to chapter 4, § 2, SoL). In their analysis of the ruling, Kjellbom and Lundberg (2018) conclude that the court ruling affirms the prioritisation of migration control over non-deported people’s access to essential welfare.

I discussed the court ruling in the light of the amended LMA with Dalia, a social worker in a Swedish municipality that is known for having a more generous interpretation of the rights of non-deported people compared to other municipalities. She told me that the ruling placed social workers in a difficult situation.

For Dalia, the LMA amendment confronted her with a professional dilemma. As a social worker, she was trained to see and meet the needs of people, whereas the new regulations implied that instead she should look away from those in need. Her reading of social workers’ approach to need – the direr your condition, the stronger your right to social assistance, no matter for how long you are in acute need – directly contradicted the logic of migration enforcement. Dalia continued, ‘We have had discussions ... what is our role? Is it to motivate them to return? No, our sole task is to alleviate suffering. But that role is in direct opposition to the work of migration authorities and the police.’ Far from all social workers took a similar position; like migration officials and border police officers, there were also social workers who maintained the differentiation in humaneness and deservingness between destitute citizens and migrant ‘others’ (see Kazemi, 2021). Yet among the social workers I interviewed, there was a widespread understanding that the hostile political climate towards people seeking protection – which was not confined to the LMA amendment but encompassed a broader range of ‘displacements’ in Swedish migration and welfare policy (Jansson-Keshavarz, Lundberg, and Obenius 2021: 319) – had changed the conditions of social work. They witnessed how people who were already in a precarious condition were made even more vulnerable, while their possibilities to assist them were becoming more limited. They also noted how responsibility for the provision of social welfare to non-deported people was gradually pushed onto civil society actors (see also European Commissioner for Human Rights, 2018); something they thought was ‘foreign’ to a welfare state. Maria was another trained social worker I interviewed, who had previously worked at one of the Swedish Migration Agency’s accommodation units for people arriving as unaccompanied minors. She told me that seeing how people were denied access to social welfare by state authorities had made her resign. When we spoke over the phone, she explained her decision to leave in the following way:

After her resignation, Maria established an NGO that provided accommodation and basic welfare services for young people whose asylum applications had been rejected – the same people she had seen state authorities abandon.

Maria’s resignation was a reaction to the discriminatory differentiation between citizens and ‘foreign’ children, which for her challenged the social imaginary of an inclusive welfare state. Concerns over what happened when the welfare state was mobilised for exclusion were also raised by migration officials. Palle, the return caseworker, told me as we finished our interview, ‘I feel like I started questioning the system. But I think, if you are going to work here, you have to buy into the system … at least to a certain extent.’ At the time we met, Palle considered resigning and applying for a doctoral position to research deportations from the perspective of non-deported people. Whether he did or not, I do not know. But the unease expressed by him, Maria, and other state officials illustrates some of the fissures and cracks in the ideology of a fair, inclusive, and egalitarian welfare state that the derecording strategy generated. The belief in the system was also shared initially among many who sought asylum in Sweden, but who were later challenged in their belief by the discrimination and disregard they experienced in their encounters with the welfare state. Their reactions to and ways of navigating the sudden changes from bureaucratic inscription to formal abandonment are important for understanding the implications of the necropolitical operation of the welfare state for non-deported people and their life chances.

Challenging deportability by moving on

In 2015, around 35,000 children arrived to seek asylum in Sweden and were registered as so-called unaccompanied minors. Many of them were Afghans. Their trajectories through the Swedish asylum system and towards deportability have become the focus of a heated public debate, and large-scale mobilisation by civil society, since they have demonstrated the discrepancy between the imaginary of a benevolent, humanitarian asylum regime, and the discriminatory, arbitrary, and harmful outcomes it generated. The treatment of this group has been discussed extensively in research detailing how the children have been exposed to racist political and societal rhetoric, framing them as tricksters, criminals, and welfare abusers (Djampour, 2018), to institutionalised suspicion, and to degrading treatment during asylum processes (Wernesjö, 2019). They have also been subjected to highly intrusive medical age assessments (Kazemi, 2021), which evoked associations to the Swedish history of racial biology, and their prospects of remaining in Sweden were constantly changed through a virtual flood of legal amendments, which amplified their uncertainty for the future (Elsrud, 2020). Recognition rates for Afghan citizens are low in Sweden, compared to other European countries, and many of the young people – who by the time they received their asylum decision had been re-classified as adults – ended up deportable, despite the volatile situation in Afghanistan (UNAMA, 2019).

Many of them refused to leave, and appealed to suspend their deportation orders. Some of them initiated demonstrations. Most famous became the sit-in protest organised in August 2017 by a group of young Afghans outside the parliament building in Stockholm, where they demanded an end to deportations to Afghanistan and the right to remain (see Khavari, 2018). During the two month-long protests, which acquired the name ‘Young in Sweden’ (Ung i Sverige), thousands of people showed up to document, support, or join the strike, including a steady stream of journalists and politicians. Ung i Sverige mobilised support from civil society organisations, including grassroots movements such as Vistårinteut (‘We cannot take it anymore’) and Stoppa utvisningarna till Afghanistan (‘Stop deportations to Afghanistan’). Like the demonstrations by people confined in the Danish deportation camps, the protests can be understood as a politics of presence; a refusal to be derecorded, a demand for participation in society, and an insistence on recognition of the violence that they had endured while in Sweden (see Khosravi, 2017b). The protesters and their support groups did not succeed in suspending deportations to Afghanistan, but they pressured politicians into proposing legislation that enabled some young people to regularise their status in Sweden temporarily. The ‘high school law’ (gymnasielagen), adopted by the parliament in June 2018, enabled unaccompanied minors who had arrived in Sweden prior to 24 November 2015 (when the ‘temporary’ law entered into force) to apply for a temporary residence permit that would enable them to finish upper secondary school (see Elsrud, 2020). It offered them a temporary respite but did not protect its holders from being deported to a country in conflict once it expired; indeed, the new law rejected the young asylum seekers’ claims to be recognised as refugees in need of protection, and instead conditioned their safety upon performances of deservingness as part of a diligent, grateful, and hard-working future workforce (Wernesjö, 2019). Therefore, critics called it ‘a prolonged deportation process rather than a real opportunity to remain in Sweden’ (Khavari, 2018), which did not resolve their condition of being in limbo.

Some young people who refused to comply with their deportation orders to Afghanistan challenged the deportation regime in other, less visible ways. Some went underground, relying on the support of friends, family, communities, and social networks. At the Swedish Migration Agency, officials were well aware of their continued presence, and regretted how their ‘futile hopes’ to find another way to regularise their status interrupted authorities’ efforts to archive their cases. Commenting on the case of non-deported Afghans, Susanna said,

Susanna argued that the young people’s hopes to remain were based on false information; but many saw the constantly changing laws, and the multiple and sometimes contradictory advice they were offered by various actors, as opportunities that were, at least, true enough for them to act upon (Eule et al., 2019). Aside from those who decided to stay and wait for novel opportunities to arise from the messy legal landscape of Swedish asylum law, Susanna also directed her frustration at the thousands of non-deported people who ‘re-escaped’ (Elsrud, 2020: 501) to other European countries in order to evade deportation from Sweden, which in theory should not be possible due to the Dublin Regulation. Indeed, since 2016, there have been reports of an increase in Dublin take-back requests from Southern European countries (see European Migration Network, 2016; Ibfelt and Skov-Jensen, 2019; Schibbye, 2019), which contradicted officials’ assumptions that most people moved along a linear, northbound journey through Europe. Swedish migration officials were unsure of whether this trend could be attributed to the new, restrictive, and deterrence oriented migration policies in Sweden, or to civil society actors, who were giving non-deported people ‘false hopes’ by encouraging them to move on. Mariam exclaimed, ‘Do they even know what Dublin is!? I don’t get what they are thinking!’

Nevertheless, people kept leaving. In 2017, Swedish news media started reporting about the growing number of young Afghans who, after their asylum applications had been rejected in Sweden, had shown up in the Swedish church in Paris (Kyrkans tidning, 2017; see also Schibbye, 2019). The Swedish church, which is at once a religious and a cultural centre, became a meeting spot for young people, mostly men, who had escaped deportation in Sweden, and who had acted upon rumours of the French asylum system being more lenient towards Afghan refugees. As more people kept arriving, the Swedish church in Paris began to offer them food, a cup of coffee, a shower, and a place to wash their clothes, since they were otherwise sleeping rough in informal tent camps or under bridges across Paris while waiting for their asylum application to be registered with French authorities. The church hired a coordinator, Klara, who also helped them translate official documents from French, arranged access to legal advice, and initiated Swedish-French language classes. I visited the Swedish church in Paris in March 2018. Around this time, Klara had recorded some two hundred young people passing through the church; some showed up only once, while others returned regularly. New people kept arriving every day.

The church café is a large open room. At one end, two Swedish-speaking women are having coffee and cake while their children are playing. Next to them, young men are sitting around the wooden tables, some of them alone, others in groups, sipping tea or coffee. Others are charging their phones, crashing on the couch upstairs, using the internet, or doing laundry. Klara runs between the laundry room downstairs and the entrance upstairs, greeting and chatting to the young men, some of whom she knows well, others whom she seems to meet for the first time. Klara explains to me,

I sit down with Abas and Rohullah. They are brothers and arrived in Sweden three years ago, Abas was a minor, Rohullah eighteen at the time. Abas explains that they both invested a lot in building a life in Sweden, learning the language, going to school, and preparing for a future there. But when their asylum applications were rejected, they did not dare to stay, and left for Paris where they heard there was an opportunity to get asylum despite Dublin. But now, they have a problem. Rohullah has been categorised as ‘normal procedure’ and awaits his first asylum interview, but Abas has been categorised as a Dublin case, and risks being deported back to Sweden. Rohullah smiles as he is telling me this story, shaking his head. But he is concerned, too, as they don’t know what to do next. They are either sleeping on the streets under dire conditions or moving around between temporary shelters. Right now, they regret their decision and would like to go back to Sweden, even though thinking of Sweden makes them angry. Abas explains,

The brothers are torn. They don’t want to leave and jeopardise Rohullah’s new-found chance to get asylum in France. If they manage to wait for eighteen months in Paris, the deadline for Abas’ Dublin transfer to Sweden will expire, and he might be able to file a fresh asylum request. But, Abas sighs, ‘The waiting is the worst, it takes over everything. Your life pauses when you don’t have papers.’

The story of Abas and Rohullah, and the estimated 4,000 other young Afghan citizens who have escaped deportation from Sweden (see Reguera and Mahmoud, 2020, though the number keeps increasing) demonstrates both the risks and opportunities arising from the uneven operation of the European deportation regime. Contrary to Mariam and other migration officials’ disbelief in the Afghans’ hopes for regularisation, their cases show how the legal condition of being rejected, illegalised, and non-deported is never fixed, and always negotiable (Wyss, 2021). The systemic illegibility of migration law renders access to rights and residency contingent upon authorities’ unpredictable legal interpretations, upon luck, and upon non-deported people’s own ability to identify and act strategically upon the offerings of the moment (Eule et al., 2019). While in theory, the Dublin Regulation and the harmonisation of asylum law across European states should prevent Rohullah and Abas from finding a second chance to get asylum, in practice, the system does indeed remain a lottery (Brekke and Brochmann, 2015; Schuster, 2011). For Afghans, asylum recognition rates vary between 6 per cent and 98 per cent between European states (ECRE, 2019), with France being among the more generous countries: in 2018, 67.5 per cent of Afghans obtained asylum in France, whereas recognition rates in Sweden were at 33 per cent.⁶ Therefore, it might not appear as surprising that in 2018, the Swedish Migration Agency reportedly recorded that there were more Afghans who re-escaped to other European countries than deportations to Afghanistan (see Elsrud, 2020).

A second important insight from the brothers’ story is the inherently transnational character of deportation regimes (Drotbohm and Hasselberg, 2015), which offer opportunities to escape, but also present risks to become trapped anew. An activist in the Danish deportation camps once noted that the Dublin Regulation effectively renders Europe an ‘open prison’ for non-deported people, either making them trapped within the confines of a state that does not want them or making them ‘stuck in transit’ (Brekke and Brochmann, 2015: 145), in a precarious condition of onward mobility (Wyss, 2019). These onward movements, and the protracted uncertainty, disruption of communities and relationships, and dire living conditions they produce, can be understood as part of the politics of exhaustion that Hasse presented in the beginning of this chapter (see Ansems de Vries and Guild, 2018). However, they also show how the deportation limbo is a site of negotiation, and is generative of new practices of control, evasion, and appropriation. If ‘luck’ and ‘chance’ remain precarious exceptions in a migration control regime set up for people to fail their migratory projects, they are also what make people endure, move on, and re-appear, contesting the efforts of states to make them disappear.

Concluding remarks

The politics of minimum welfare, of exhaustion, and of derecording enable Swedish state authorities to govern non-deported people using minimum force. Through the logic of voluntariness and responsibilisation, non-deported people are made responsible for their vulnerable condition; it is up to them, the narrative goes, to inform themselves of their condition, and to make the ‘right’ decisions and comply with the attempts of the state to deport them. Officially, these measures all fall under the category of what constitutes ‘voluntary’ return measures, which is Swedish authorities’ preferred way of enforcing deportations. Such discourses of benevolence and free choice function to conceal the structural violence of making people destitute, and the symbolic and legal violence of compelling non-deported people to act under the false pretence of freedom. The violence therefore goes unrecognised, for ‘violence that occurs gradually and out of sight […] is typically not viewed as violence at all’ (Davies and Isakjee, 2019: 214). It simultaneously makes non-deported migrants disappear, from bureaucratic registries, and sometimes, from the country. At the same time, it enables state violence to disappear, making it difficult to trace, other than through its effects.

The hiding away of these indirect forms of state violence is made possible by the devaluation of non-deported people’s lives. The harms caused by this violence are unevenly distributed among different groups. Without comparing the direness of their conditions, a young man in good health might have higher chances to survive without formal access to welfare than a woman of older age who is dying of cancer. A family with young children might find it more difficult to make the decision to move on within Europe, considering the risk to end up destitute and without alternative support networks (although some do, as this option is preferable to deportation). Access to formal or informal support is also determined through gendered and racialised notions of deservingness; in Sweden, there were specialised (and state-funded) NGOs working to support primarily young Afghans who had been through the asylum process, whereas there was significantly less support available to people like Samir, the detained North African man who appeared in Chapter 3.

To some extent, the purported failure to protect those in most dire need contradicted state officials’ imaginary of a humane and egalitarian welfare state. The ‘displacements’ (Jansson-Keshavarz et al., 2021) in the support mandate for social workers, in particular, caused many to question the conditions of their job, some to protest, and some, like Maria, to resign. While the welfare state as such is a bordering device, and while limited access to welfare is a common strategy of deterrence among European states, I concur with Davies et al. (2017; see also Khosravi, 2010) that the institutionalised and organised nature of abandonment in wealthy, bureaucratised welfare states makes acutely visible the political prioritisation of some lives over others. Moreover, the radical forms of exclusion it generates need to be seen in the context of a civil society unprepared to fill the ‘gaps’ in welfare provisions. The examples I have provided from Denmark and Sweden indicate the importance of abandoning simplistic understandings of welfare states and bureaucracies as egalitarian institutions and neutral distributors of ‘universal’ rights.

Finally, the effects of the deterrence strategies of the Nordic welfare states extend beyond their borders. The significant number of people who have left both Denmark and Sweden after a final rejection on their asylum application testify to how responsibility for providing protection, for processing applications, and for enforcing deportations are pushed onto countries in Southern Europe. Their trajectories illustrate the transnational expansion of the deportation limbo, and how it generates a proliferation of sites of enforcement, and of contestation (Sanchez Boe, forthcoming; Schmid-Scott, 2018). They also demonstrate the inadequacy of limiting an analysis of deportation policies within the confines of a single state apparatus or legal framework. We need, therefore, to consider how deportation regimes expand: inwards, outwards, and onwards.