Walled in Not Walled Off?

Sweden has reformed its approach to child detention, introducing youth wings within regular adult prison complexes to house 15- to 17-year-olds convicted and sentenced to detention. This reform forms part of a broader scheme to tackle organised crime, in the course of which criminal groups in Sweden have increasingly instrumentalised children to carry out offences. Previously, children for whom deprivation of liberty was unavoidable were accommodated not within the general prison service but in separate residential social care homes administered by the National Board of Institutional Care (Statens Institutionsstyrelse). The newly introduced youth wings are administered by the Swedish prison service (Kriminalvården). Initially, the reform included lowering the age of criminal responsibility to 13, consequently allowing placement of children from that age in adult prison complexes for up to 18 years. This drew stark criticism even from within the prison administration, with the director of Kumla prison stating that children had no place in prison. While the lowering of the age of criminal responsibility was intermittently halted in June, in August of 2026 the Swedish Parliament lowered the age of criminal responsibility to 14. This policy shift raises questions as to the scope of separation of children and adult detainees under Article 37(c) of the Convention on the Rights of the Child (CRC).

The Meaning of Separation in Article 37(c) CRC

Article 37 CRC forms the basis for the administration of criminal sanctions against children. The overarching principle, enshrined in Article 37(b) CRC, is that deprivation of liberty as a form of punishment ought to be the ultima ratio. The Committee on the Rights of the Child (the Committee) describes the effects of deprivation of liberty as being detrimental to the health and safety of children, increasing recidivism, and diminishing their prospects of reintegration. Article 37 CRC then provides safeguarding standards for children who are exceptionally deprived of their liberty as part of a criminal sanction. According to Article 37(c) CRC, children must be separated from adults in detention unless it is not in the best interests of the child. The Swedish reform brings into focus whether the separation requirement under Article 37(c) CRC is sufficiently met by a formal, spatial exclusion of adults from child detention or whether it also requires protection against the risks, including labelling and discursive effects, that result from the child’s placement in proximity to adult detainees.

The Drafting of Article 37(c) CRC

There was a lack of consensus during the drafting of Article 37 CRC and no specific debate on the institutional requirements of separation. The United States sought to qualify the provision on the separation of children from adults by including an exception for cases where the law determines that a child is to be treated as an adult. Algeria objected to the US’s proposition, stating that “the purpose of the Convention would be defeated”. Strikingly, almost every reservation states parties have made against Article 37 CRC concerns the separation requirement. Yet, the content of these reservations is not uniform. The reservations do not show a shared state understanding of what form of institutional separation Article 37(c) CRC requires. Rather, they reflect a divided and inconsistent practice ranging from rejection of strict separation to its qualified acceptance. The Committee has continuously criticised the effects of these reservations but has not pronounced on the institutional design requirements flowing from Article 37(c) CRC.

Separation in Light of the Purpose of the CRC

As a starting point, separation in the narrow formal sense of mere exclusion of adults from child detention can be understood as a minimum requirement of institutional design. First, because of the punitive design of adult correctional systems, which is at odds with a child-centred, needs-based approach to youth justice, which must not rely on punishment as the primary means of reintegration. Secondly, because of the risk of contact with adult offenders, which exposes children to potentially harmful influences, further exploitation and abuse. The two explanations locate the risk within the adult facility in the form of punitive design and social environment.

The primary rationale is that the placement of children in the adult prison population is detrimental both to the child’s safety and to their prospects of reintegration. Throughout, the Convention recognises that children are in a state of ongoing development. Article 37(c) CRC provides that a child in detention shall be treated “in a manner which takes into account the needs of persons of his or her age”. This did not lead to an outright rejection of child detention in the Convention, but the rationale survives structurally in Article 37(b)’s last-resort and shortest-appropriate-period requirements. Both give effect to the overarching best-interest principle in Article 3(1) CRC. Any sanction imposed upon a child must therefore show regard for the developmental potential of children and take it as a basis for reintegration. This bears direct institutional consequences. Not only as to where a child is placed but also as to how the facility itself is structured. The United Nations Rules for the Protection of Juveniles Deprived of their Liberty, referenced throughout the Committee’s practice, recognise this tie between developmental need and institutional design.

Article 37(c) CRC therefore cannot be read in isolation from the Convention’s developmental and reintegration-oriented approach. This approach is, however, not a one-way street. Article 37(c) allows non-separation for children, and an expansion of separation to those having reached adulthood in detention where it is in the best interest of the child. If best interests are at the core of both outer cases alike, the protective function of separation is not defined by age as such, but rather by its role as a means to guard against a risk profile for which age is one marker.

In light of the best-interests right as an “interpretative legal principle”,  the essential content of separation lies not in the formal spatial exclusion of adults from child detention. Rather, it lies in the protective and developmental rationale that spatial exclusion is aimed at securing – namely, shielding children from punitive institutional design and from a harmful social environment, while simultaneously enabling the reintegrative purpose of any sanction against children. Physical separation is accordingly best understood as the archetypal means of achieving this protection, not as an end in itself. Article 37(c) CRC itself treats formal separation as displaceable where it fails to provide the necessary protection under best-interest considerations. The assessment of compatibility with Article 37(c) CRC is therefore not limited to whether children are detained apart from adults. It must ask whether the institutional design excludes the adverse effects of adult detention and whether it affirmatively supports the reintegrative and developmental approach of the CRC to criminal sanctions against children.

Discourse and Labelling – Risks of Juvenile Wings in Adult Prison

If Article 37(c) CRC refers to the separation of children from risks inherent in adult detention, it remains to be addressed whether juvenile wings within adult prison complexes are compatible with the separation requirement.

Discourse Analysis – What the Swedish Reform Reveals

Sundhall and Hillén’s discourse-analytic study of the Swedish policy shows that the proposal to replace social care settings with “youth prisons” run by the (adult) Correctional Service makes “invisible the power imbalance between children and adults”. Excluding language such as “child” or “care” in the discourse only where children are deviating from norms arguably puts the burden upon the child to show that they deserve to be considered vulnerable. This is a notable inversion of the Committee’s approach to structurally comparable cases. Children associated with non-state armed groups, who are also both recruited and instrumentalised, are to be “treated primarily as victims” under international law, with prosecution and detention reserved as a last resort. No equivalent presumption is extended to children instrumentalised by organised crime in the policy. While the policy’s own drafters concede that these children are affected by the same conflicts, the policy resolves this ambiguity with a structural offender-framing of children. The best interest of the child is regrettably not the essential concern here, but rather the creation of an impression that the government is cracking down on crime. The integration of juvenile detention into adult prison complexes works in a similar manner, not semantically, but structurally. It is itself an act of categorisation, classifying child offenders as though they were adults.

The Consequence of Shifting the Discourse

Placing children in adult correctional structures, in spite of spatial separation within buildings, has a discursive effect on how detained children are perceived. However, it may also affect self-identification.  Identity can be understood as produced through the societal reaction to a person’s formal categorisation, rather than solely through behavioural learning or contact with other offenders. This process of “labelling, stereotyping, separating, status loss, and discrimination” operates through structural and institutional pathways such as placing children in a physical complex reserved for the most severe state response to norm deviation. The effect of this is not limited to external perception and stigmatisation or discrimination. It may impact the self-identity of children placed therein so that children affirm the stigma rather than reject it. This effect on self-identification can in turn hinder reintegration of children: firstly, by perpetuating affiliation with delinquency as a framework of belonging; and secondly, by perpetuating external stigma, creating barriers to reintegration. A formally imposed criminal status can then, especially in the context of organised crime, which the policy is supposed to tackle, become a source of standing within a peer economy that already rewards proximity to violence and criminal competence. A study of criminally involved youth in Sweden documents precisely this logic in relation to guns, which function as markers of “violence capital” and reputation among young offenders. An adult prison, even where its youth wing is nominally separate, is liable to operate as an equivalent marker.

Against these considerations, the separation requirement in Article 37(c) CRC is better understood as a risk-proximity standard than a formal spatial standard. The latter alone is not capable of excluding the risks associated with adult detention, and integrated youth wings in adult prison complexes are ill-suited to excluding such proximity-related risks.

Conclusion

The Swedish policy reform deviates substantially from the rationale of Article 37(c) CRC even as it preserves the formal appearance of compliance. The discourse on child exploitation by organised crime and the institutional reforms are two expressions of the same underlying shift towards reclassifying the child as a perpetrator rather than a victim. This emerging picture of the child as the problem, the criminal who is being managed and not pampered in social care institutions, may be resonating in the political climate in Sweden, but it sidelines the best interests of children in favour of creating the impression that the government is cracking down on crime. The child becomes an object of control rather than protection. A narrow reading of Article 37(c) CRC cannot alone exclude the risk factors the provision is meant to guard against because it only asks whether children and adults occupy separate premises. It does not ask whether the risks separation is designed to protect against are in substance mitigated or excluded. Separation understood as a risk-proximity standard for institutional design can fill this gap.

The “Bofaxe” series appears as part of a collaboration between the IFHV and Völkerrechtsblog.