September 11, 2026
By Sophie Bols
The ECtHR’s judgment in D.M. v Sweden has been well received by scholars and practitioners, alongside other recent migration-related judgments, in which the Court has issued well‑reasoned and protective decisions amid intense political contestation. The judgment provides important safeguards at a moment when several member states have begun deporting, or openly discussing deporting, Afghan nationals to Taliban‑ruled Afghanistan. While this post was being written, representatives of the Taliban were invited to Brussels to discuss forced returns. Only one week earlier, the European Parliament adopted the Return Regulation, while right-wing MEPs celebrated this by chanting “send them back”.
Against this backdrop, the Court’s unanimous finding that deporting an Afghan national of Hazara ethnicity would violate Article 3 of the Convention is a significant and welcome counterpoint. D.M. v Sweden is also the first judgment concerning a deportation to Afghanistan since the Taliban takeover in 2021. This post examines the Court’s reasoning more closely, focusing on its treatment of Hazara ethnicity, its engagement with “westernisation” as a risk‑enhancing factor (including whether this can realistically be concealed), the perception of apostasy, and the cumulative risk assessment that ultimately led the Court to find a violation of Article 3.
Mr. D.M., an Afghan national of Hazara ethnicity, arrived in Sweden in 2015. Over the years, he submitted multiple asylum applications, invoking a land dispute, his participation in a demonstration, his conversion to Christianity, his apostasy from Islam, and his Hazara identity. His claims were consistently rejected. After the Taliban takeover in August 2021, he lodged a new application, arguing that the changed situation in Afghanistan placed him at risk. The Migration Agency again rejected his claim, maintaining earlier credibility findings and concluding that neither his Hazara ethnicity, nor his alleged “westernisation”, nor the Taliban takeover established a personal risk. The Migration Court upheld the decision.
On 3 November 2023, the ECtHR applied Rule 39 interim measures, staying his removal. The case proceeded to judgment in March 2026.
The applicant complained that his removal would expose him to a real risk of treatment under Article 3 and Article 2, and that the domestic proceedings had not met the standards required under Articles 2, 3 and 13. The Court reiterated that it is “the master of the characterisation” of the facts and examined the case solely under Article 3 (para 130), demonstrating what has been called the Court’s “exclusive approach” to Article 2 in expulsion cases.
The Swedish government argued that the Court should not substitute its assessment of the facts for that of the domestic authorities. The Court agreed and reiterated its principle that national authorities are best placed to assess credibility (para 153). It therefore did not disturb the domestic credibility findings, instead basing its judgment on factors the Swedish authorities had accepted or left undisputed (para 184). Given this evidentiary background, the Court may have judged that the applicant did not present a sufficiently concrete threat to life to warrant a separate Article 2 analysis. Moreover, Article 3 has generally historically overshadowed Article 2 in non‑refoulement jurisprudence.
The Court emphasised that where the applicant has not yet been deported, the material point in time for assessing Article 3 risk is the time of the Court’s consideration. It therefore conducted a full ex nunc assessment, relying on extensive expert and country information, including UNHCR, UNAMA, the UN Special Rapporteur, EUAA guidance, and reports from NGOs and national authorities (paras 78 – 80, 82, 89, 95, 120 – 127).
The Court held that while the security situation in Afghanistan remains serious and fragile, the general situation alone is not sufficient to conclude that all returns would lead to a violation of Article 3 (paras 162 – 166). It recognised that Hazaras face widespread discrimination and are targeted for deadly attacks by the ISKP (Islamic State Khorasan Province), and that the Taliban have failed to protect Hazara communities (paras 167 – 176). Yet it concluded that Hazaras cannot be considered a group systematically exposed to ill‑treatment reaching the Article 3 threshold (para 177), while simultaneously emphasising that Hazara ethnicity may significantly increase risk when combined with other factors.
The core of the judgment lies in the cumulative assessment of D.M.’s individual circumstances such as his Hazara ethnicity and area of origin, his long stay in Sweden and “westernisation”, and behaviour perceived as transgressing religious and moral norms. The Court held that the domestic authorities failed to conduct a proper cumulative assessment and concluded that these factors taken together created a real risk of ill‑treatment. Deportation would therefore breach Article 3.
The applicant argued that he would face ill‑treatment owing to his Hazara ethnicity (para 167). Hazaras form a historically marginalised minority, making up around 15% of the Afghan population. They have long occupied the bottom of Afghanistan’s ethnic hierarchy, are predominantly Shia Muslims in a majority‑Sunni country, and have faced systemic discrimination on both religious and racial grounds. The Court had previously held (here and here) that this did not reach the Article 3 threshold, but that assessment pre‑dated the Taliban takeover (para 168). As D.M. v Sweden is the first Afghanistan deportation case since then, it offered an opportunity to revisit this position.
In its non‑refoulement case law, the Court distinguishes between different types of risk: first, generalised violence, second, membership of a group systematically exposed to ill‑treatment, and third, individual circumstances creating a real risk. In D.M. v Sweden, it declined to place Hazaras in the second category. Hazara applicants therefore cannot rely on group membership alone and must demonstrate additional individual risk factors.
The Court nevertheless recognised that Hazaras face widespread discrimination and deadly attacks by ISKP, and that the Taliban have failed to protect them (para 175). Hostility towards Hazaras is rooted in ethnicity and religion, reinforced by perceptions that they “benefitted too much” under the previous government (para 172) and have embraced “a culture at odds with the Taliban’s definition of Islam” (para 91).
The conclusion that Hazaras are not systematically exposed to Article 3‑level ill‑treatment may have ambivalent domestic consequences. On the one hand, the judgment strengthens protection against deportation as it insists on a cumulative risk assessment and explicitly recognises Hazara ethnicity as a serious risk‑enhancing factor. On the other hand, domestic authorities might seize upon the Court’s statement to resist group‑based presumptions of risk and stick to narrow, individualised assessments. Asylum authorities might even also be tempted to use it to argue against group‑based persecution, but any impact on asylum decision‑making must be carefully nuanced. The ECtHR does not concern itself directly with obligations under the Refugee Convention or EU asylum law, and Article 3 ECHR has a different threshold and object than the Refugee Convention. Under international refugee law, individuals who belong to a group facing systematic persecution need not show individual targeting. Across Europe, protection has been granted on the basis of group membership alone where country information demonstrates a systematic practice of persecution.
This approach was confirmed by the CJEU in 2024 in joined cases AH and FN – the “Afghan women” cases – where the Court, in line with what had already been recognised in many national practices, held that the cumulative discriminatory measures imposed on Afghan women amount to “acts of persecution” under the Qualification Directive and that an individualised assessment is unnecessary.
In the current political climate, selective citation of D.M. v Sweden is not unthinkable. Politicians or asylum authorities may use the Court’s statement to argue against group‑based recognition of Hazara Afghans or to justify forced‑return initiatives. The political sensitivity of the case was also reflected in the Court’s press Q&A, where it emphasised that this judgment does not mean that all deportations to Afghanistan breach the Convention.
Interestingly, the Court appears to anticipate this risk by stating that its finding is “without prejudice to whether the acts, or accumulation of acts, against Hazaras could amount to persecution as defined by the Refugee Convention and the Qualification Directive” (para 177). It also cites the UK Home Office’s position that ethnic Hazaras, mostly Shia Muslims, continue to face abuses by the Taliban and are generally likely to face a real risk of persecution (para 117).
A faithful engagement with the Court’s reasoning would require more careful examination of Hazara applicants’ circumstances. Properly understood, the judgment strengthens protection against expulsion for Hazara Afghans.
In addition to his Hazara ethnicity and his apostasy from Islam, the applicant relied on his so-called “westernisation”. While the concept is not entirely new, it has become increasingly relevant in Afghan protection claims since the Taliban’s return to power. Scholars have warned that without clarification, the concept risks reinforcing stereotypes or grounding protection in perceived assimilation into “European values”.
Domestic authorities often assess westernisation through their own cultural lens. In D.M. v Sweden, the Court took a different approach, relying on the applicant’s written evidence, scholarly arguments, and country information. The applicant submitted an opinion by Professor Emeritus William Maley, who explained that individuals who spend formative years abroad often return with dated vocabulary, altered body language, gestures, and mannerisms, which are traits they may not even themselves notice. In Afghanistan, such individuals may attract the derogatory label “gharbzadeh” (“child of the West”), triggering prejudices and animosities that place them at real risk (para 135).
Importantly, the Court assessed westernisation through the eyes of Afghan society and potential actors of persecution, not through the applicant’s internal identity. It approached the applicant’s alleged conversion and apostasy from Islam in a similar way, focusing on perception rather than on his subjective religious experience. Although the Court did not disturb the domestic authorities’ credibility findings regarding the sincerity of the applicant’s conversion, it held that the risks he may face when perceived as a convert or apostate must be taken into consideration.
The Court noted reports indicating that people perceived as having transgressed religious, moral, or societal norms (even when abroad), and those seen as influenced by foreign values may face ill‑treatment from the Taliban, but also from relatives, neighbours, or other actors (para 191). It emphasised several aspects of the applicant’s profile, such as his arrival in Sweden as a teenager or young adult, his lack of recent experience in Afghanistan, his negative attitude towards Islam, his criticism of Afghan society, his long‑standing non‑practice of Muslim rituals, his lack of knowledge of prayers, and the fact that he spoke Dari with a slight Swedish accent. It also noted his baptism, church attendance, and participation in church activities. Holding that this behaviour is likely to be perceived as violating religious and moral norms (para 192). The Court further considered the risk of inadvertently violating the Law on the Promotion of Virtue and Prevention of Vice (para 194), noting troubling indications that the law is implemented even more strictly in areas where ethnic and religious minorities are based, including Mazar‑e Sharif (para 196).
The Swedish Migration Agency and Migration Court held that the applicant could and should adapt to Afghan customs (para 185), suggesting that outward conformity to Muslim rituals and Taliban‑mandated social codes would suffice. They also argued that westernisation was not a fundamental, immutable characteristic and could be concealed.
The Court disagreed but did not weigh in on whether westernisation is a core aspect of one’s identity that cannot be required to be hidden. Instead, it focused on the practical impossibility of concealment in Afghanistan.
By contrast, the Court has held in its previous case law that certain characteristics, such as sexual orientation, constitute fundamental facets of an individual’s identity that cannot be required to be concealed. It also referred to the CJEU’s case law, specifically Y and Z (C‑71/11 and C‑99/11), where concealment of religious belief could not be demanded, and K, L v Staatssecretaris (C‑646/21), the “westernised women case”, in which the CJEU held that Afghan women could not be expected to restrain their expression of gender‑equality values to avoid persecution.
In D.M. v Sweden, however, the Court sidestepped the question of immutability and instead emphasised the practical impossibility of “playing the game”. It relied on Sufi and Elmi, where it held that a Somali with no recent experience of Somalia would be unlikely to navigate a repressive environment successfully, and that this difficulty would be even greater for individuals who had become “westernised”.
In the present case, the Court stated that even assuming concealment could be required, it was not convinced the applicant could adequately hide these aspects of his life or that concealment would obviate the risks (para 193). It emphasised the nature of the current repressive regime, the strict rules governing almost every aspect of life, and the active monitoring by the Taliban and local communities. The Court ultimately held that what matters is the real‑world likelihood of exposure to ill‑treatment, not the theoretical possibility of behavioural adjustment.
Although the Court concluded that Hazaras cannot be regarded as a group systematically exposed to ill‑treatment reaching the Article 3 threshold (para 195), it emphasised that individual factors which may appear insufficient when assessed separately can, when taken together and considered in light of the general situation in Afghanistan, give rise to a real risk. The Court found no indication of such a cumulative assessment in the domestic decisions (para 197).
The Swedish authorities had assessed each alleged risk in isolation. The ECtHR criticised this fragmented approach and considered all of the applicant’s personal circumstances cumulatively. Taken together, and in light of the broader human rights situation in Afghanistan, these factors created a real risk of ill‑treatment upon return (para 197).
The emphasis on cumulative reasoning is particularly important in the current political climate, where member states seek to argue that Afghanistan is “safe” for certain categories of returnees. The judgment underscores that even where no generalised risk exists, and even where individual factors may not independently reach the Article 3 threshold, their combined effect may still warrant protection.
D.M. v Sweden is a strong signal regarding possible deportations to Afghanistan in the post‑Taliban takeover context and at a time when several member states are exploring return arrangements with the Taliban. While the Court did not recognise Hazaras as a group systematically exposed to Article 3‑level ill‑treatment, it emphasised that Hazara ethnicity is a serious risk‑enhancing factor that must be assessed cumulatively with other relevant elements. The Court’s engagement with “westernisation” offers welcome guidance on how such claims should be understood and assessed, focusing on perception in the country of return rather than assimilation or internal processes. The judgment strengthens protection against expulsion and underscores the continuing relevance of Article 3 as a safeguard in an increasingly hostile political climate.
Strasbourg Observers is an academic blog that discusses recent developments at the European Court of Human Rights. The blog is run as an inter-university collaboration amongst the Dutch-language universities of Belgium. Associated research groups are the Human Rights Centre at Ghent University, the Centre for Government and Law at Hasselt University, the Leuven Centre for Public Law at KU Leuven, the Law and Development Research Group at the University of Antwerp and the Vakgroep Publiekrecht at the Vrije Universiteit Brussel.
Strasbourg Observers would not be possible without the valued work of countless guest bloggers. Visit the “Contact” page on our website to find out how to become one!
