Kurt t. Oostenrijk (EHRM, 62903/15) – Accounting for the particular context of domestic violence

Overview

1. Within a year, the European Court of Human Rights has delivered six judgments in cases where domestic violence has been the core issue: Kurt v. Austria (a Grand Chamber judgment);[1] Tkhelidze v. Georgia;[2] Tunikova and Others v. Russia;[3] A and B v. Georgia[4], Y and Others v. Bulgaria[5], and Landi v. Italy.[6] This case note focuses on the first five cases, which confirmed the need to adapt the Osman test to ‘the particular context of domestic violence’ as previously established in Opuz,[7] Volodina,[8] and Talpis[9]. They aim to provide greater clarity with respect to key aspects of the test, specifically in relation to the relevance of the ‘context of domestic violence’ for the assessment of lethal risk, the impact of domestic violence on children, and the use of protection orders and pre-trial detention as part of the States’ ‘toolkit’ to prevent violence from happening. The judgments also suggest an improved understanding, on the part of the Court, of the interconnections that exist between gender-based discrimination and violence. That being said, insufficient recognition of the impact of additional grounds of discrimination on the experiences and manifestation of violence may hinder the effective application in the future of these recently established standards.

2. Kurt v. Austria, the first Grand Chamber decision involving domestic violence issues, concerned the murder of an 8-year-old boy who was shot on school premises in front of his younger sister by their father, who then committed suicide. In this case, the father was on probation and had protection orders made against him, after being previously convicted for domestic violence offences. A new criminal procedure had been initiated after he committed further violent acts against the mother when she filed a claim for divorce. In a much critiqued and divisive judgment the Grand Chamber rejected the applicant’s claim of a violation to Article 2 of the Convention on account of the authorities’ failure to ensure the protection of her son’s life from his violent father.

3. The case of Tkhelidze concerns the murder of a university professor by her ex-husband, after enduring domestic violence for a long period and forcing her to move to another city. The victim in this case made numerous reports of the violence and threats against her and her child to the police, pleading for State protection. In all cases, the police limited themselves to recording the statements without issuing any protection measures. She was shot by her ex-husband in her workplace.

4. In Tunikova and Others, four applicants suffered physical violence at the hands of their (former) partners, which was documented in medical records and in police reports. Ms Tunikova suffered a concussion, bruises and abrasions; Ms Gershman was attacked multiple times inside and outside her home; Ms Petrakova reported assaults over seven years; and Ms Gracheva’s was mutilated and ended up suffering from life-long physical disabilities.

5. A and B v. Georgia is a case concerning the murder of a woman by her (ex)partner, despite multiple reports to the police and the prosecution office. In this case the Court found that the abuser’s status as a police officer granted him further impunity, as evidenced by the reluctance of the police to investigate the allegations of violence and the emphasis the prosecutorial office placed on the compromised testimony of the police. While the abuser was sentenced to prison and the next of kin received compensation, no criminal investigation was initiated against the police agents for their inactivity and negligence.

6. Lastly, Y and Others v. Bulgaria concerns the murder of a woman by her husband, who shot her in a café after an escalation of threats and violence when she asked him for divorce. In this case, the authorities failed to properly assess the risk for her life and showed negligence in the enforcement of the protection order in place against the abuser, despite the victim’s reports to the police about the his continuing breaches of the order and his legal possession of a firearm.

7. The standard of protection offered by each of the respondent States varied significantly in these four cases, ranging from the complete absence of a legislative framework to address domestic violence and protect victims (Tunikova, and previously, Volodina) to the negligence of public agents when applying existing frameworks (Tkhelidze, A and B and Y and Others) and a largely developed system of response and protection against violence, including trained professionals (Kurt). The diverse settings and conditions of the cases allows for a more comprehensive and critical understanding of what ‘the context of domestic violence’ means for the Court.

The obligation to establish a legal framework

8. In these cases, the Court has confirmed that the States’ positive obligations under Article 2 and 3 of the Convention comprise, firstly, an obligation to put in place a legislative and regulatory framework of protection.[10] Such a legislative and administrative framework needs to be designed to provide effective deterrence against threats to the right of life. In cases involving acts of domestic violence, the Court has clarified that an effective framework would usually require the adoption of criminal law measures, including the criminalisation of acts of violence within the family by providing effective, proportionate, and dissuasive sanctions. This can be achieved either by adopting a separate offence or as an aggravating element of other offences.[11]

9. The justifications for the requirement of specific criminalization, even when acts of domestic violence could fall under the general existing provisions, are: (1) to avoid disparate responses (administrative or criminal law) that depend upon the severity of the acts in question; and (2), to treat domestic violence as a single course of conduct that results in a continuum of domestic violence, instead of separate sanctionable instances.[12] This second aspect gives testimony of the Court’s growing understanding of the particular dynamics of domestic violence. In addition, the effective criminal law provisions must be reinforced by law-enforcement machinery for the prevention, suppression, and punishment of breaches of such provisions.[13]

10. Furthermore, the Court has indicated that, in addition to criminalization, the domestic legal framework must adequately afford protection to victims against acts of violence perpetrated by private individuals. Authorities should be able to choose from a ‘toolbox’ of legal and operational measures they consider adequate and proportionate to the level of risk in each concrete case.[14] These measures should allow the authorities to ‘avoid a dangerous situation as quickly as possible’.[15] This includes immediate protection measures (e.g. restraining orders, protection orders or safety orders) that are aimed at preventing the recurrence of violence and protecting the victim by requiring the perpetrator to leave a shared residence and refrain from approaching or contacting the victim.[16] Whether the toolbox is comprehensive enough will be determined based on the protection (potentially) afforded in a specific case. All judgments confirm that protection orders are an essential part of this toolkit. The Court thereby emphasises that most European countries have implemented protection order regimes, largely as a consequence of the obligation on EU Member States to implement the EU Victims’ Rights Directive[17] and, amongst Council of Europe members, the Istanbul Convention.[18] There are a few exceptions to this general tendency, including Russia, which the Court commented upon by requiring the Russian government to introduce protection orders without delay.[19]

11. The Court gave greater detail as to what these protection orders ought to entail in Tunikova. They should: (1) be available independently of any other legal proceedings; (2) be granted taking into account the evidence victims can reasonably offer instead of requiring a criminal standard of proof; (3) indicate a specified minimum distance from the victim for the perpetrator to respect at all times; and (4), prohibit the perpetrator from attempting to contact the victim in any way, offline or online (‘no contact order’).[20] Authorities should also monitor compliance with the terms of the protection order ‘rigorously and continually’, criminalise the failure to comply, and impose dissuasive and deterrent sanctions for the lack of compliance. The characteristics identified by the Court confirm their evolving understanding of the requirements of protection, accommodate the findings of empirical studies and echo the calls of other specialised bodies such as the independent expert body responsible for monitoring the implementation of the Istanbul Convention (GREVIO).

12. In addition to protection orders, the Court has noted in these cases that an effective preventive response to domestic violence generally requires coordination among multiple authorities,[21] and in this sense, the Court has found that risk management plans, coordinated support services and perpetrator programmes all contribute to enhance such protection.[22] That said, in Tunikova, the Court was critical that most available measures targeted the perpetrator (re-education and rehabilitation) rather than focused on protecting the victim.[23] A balanced approach tailored to the specific case seems to be the proper response.

13. The case of Kurt considered the issue of whether pre-trial detention ought to be one of the tools available in the ‘toolbox’. The Court recalled that Article 5 of the Convention imposes particular constraints on measures that entail a deprivation of liberty. It clarified that preventive detention can be used to ensure that a criminal offence is not committed in the imminent future, as long as the rules governing the use of such a measure are concrete and specific, for instance, because specific measures have been ordered and violated.[24] Moreover, the Court considered that pre-trial detention can be used as a preventive measure if there is a reasonable suspicion of guilt concerning an existing offence for which criminal proceedings are pending.[25]

14. While these four cases enumerate the main aspects that the legislative frameworks should provide, States’ positive obligations to act will be determined by the assessment of the risk of violence. It is here that a comprehensive understanding of the ‘particular context of domestic violence’ becomes particularly relevant.

The obligation to prevent the known risk of ill-treatment

15. Context sensitive assessment of the Osman test. The Court has emphasised that in domestic violence cases a risk of a real and immediate threat must be assessed, taking due account of the particular context in which it takes place. This means that in assessing the ‘immediacy’ of the risk, the specific features of domestic violence cases, such as consecutive cycles of violence and the increase in frequency and intensity over time, must be taken into account.[26] Moreover, the Court has recognised that ‘where there is a lasting situation of domestic violence, there can hardly be any doubt about the immediacy of the danger posed to the victim’.[27] In such situations, it is not only a question of an obligation to afford general protection to society, but above all to take account of the recurrence of successive episodes of violence within a family.[28] Establishing a two-step process, the Court clarified that the purpose of conducting any risk assessment is: (1) to identify the risk; and (2), to provide coordinated and effective measures of protection and support to the victims.  Furthermore, the Court held that in situations where several persons are directly or indirectly affected by domestic violence, any risk assessment must systematically identify and address all the potential victims. This may result in different identified levels of risk and, consequently, require responsible government institutions to produce tailored responses.[29]

16. Risk assessment. The four recent judgments further have clarified the requirements for the risk assessment that must be performed in the domestic violence context, providing detailed guidance for future cases. Firstly, authorities have a duty to undertake an ‘autonomous’, ‘proactive’ and ‘comprehensive’ risk assessment of the victim’s treatment in order to be able to determine whether it is contrary to Articles 2 and/or 3. In this regard, the Court has recommended using standardised risk assessment tools and checklists. Secondly, the Court has explained that, while the victim’s perception of risk should be considered in the assessment, authorities must also carry out their own assessment, including by collecting information on the factors that can increase the risk of violence in a case, for instance, from other State agencies. Moreover, victims should be kept informed of the outcome of the risk assessment. Finally, the assessment should be documented in some form and communicated to those who come into regular contact with the individuals at risk.[30]

17. The Court has also outlined the kinds of factors State authorities ought to be aware of when assessing the risk of recurrent violence, namely the perpetrator’s history of violent behavior; the breach of a protection order; an escalation of violence; access to weapons; and, the existence of repeated pleas for assistance by the victims (emergency calls, formal complaints and petitions to the police).[31] In these cases, the Court has noted perpetrators gambling addiction (Kurt), mental instability (Kurt, Tkhelidze), pathological jealousy (Tkhelidze, Tunikova), economic dependency (Kurt), unemployment (Kurt, Tkhelidze), and drug and alcohol abuse (Tkhelidze) as factors that should suggest an increased risk. The Court, however, remained silent about the factors increasing victims’ vulnerabilities. While the exclusive focus on the perpetrator is a limitation of the Court’s approach, the emphasis on ‘factors’ without sufficient attention to structural conditions is also problematic, as discussed in the next section.

18. Once the risk has been identified, the authorities must adopt, as diligently and as quickly as possible, operational preventive and protective measures that are adequate and proportionate to the risk.[32] The assessment of the ‘adequacy and proportionality of the operational and preventative measures’ revealed violations of Articles 2 and 3 of the Convention in all cases with the exception to Kurt, in which the Court did not actively consider what it means to take into account the specific context and dynamics of domestic violence.

19. As mentioned, the Court has emphasised that protection orders are essential to counter the risk of violence. In the case of Austria (the respondent State in Kurt), the protection order system met most of the criteria established by the Court, since it was indeed independent of other legal proceedings and offered a combination of police emergency orders and longer-term civil protection orders, in addition to the provision of support services. This ‘Austrian model’ of protection orders has largely been welcomed by different human rights bodies and inspired similar systems in multiple states.[33] That being said, two shortcomings in the system are identified in Kurt.

20. Firstly, drawing from the requirements of article 52 and 53 of the Istanbul Convention and the observations made by GREVIO, both in their monitoring of Austria[34] and as third party submission to the case, the Court recognised that limiting the scope of protection orders to the home and other spheres regularly attended by the victims, while at the same time excluding children’s schools unless victims made a specific request before the judge, provided insufficient protection.[35] In Kurt, the Court acknowledged that perpetrators may exert violence on the children of the partner, including deadly violence, as ‘the ultimate form of punishment against their partner’.[36] The Court also confirmed that ‘interference by the authorities with the alleged perpetrator’s private and family life may be necessary to protect the life and rights of the victims and to prevent criminal acts directed against their lives or health’.[37] Furthermore, it considered that protection orders should take this aspect into account by granting protection not only within the home (the ‘place’ where domestic violence is considered to take place) but also outside, particularly in schools and other places the children in question regularly access. Nevertheless, regardless of these general issues, the Court found in Kurt that Austria had not violated its obligations under Article 2 since ‘the risk assessment did not indicate a real and immediate lethality risk to the applicant’s son’.[38]

21. The second shortcoming relates to the use of (pre-trial) detention to prevent the risk of retaliation. Here there appears to be some contradiction in the Court’s reasoning. Arrest of the abuser was considered as one of the possible measures that the Bulgarian authorities could have taken in Y and Others to prevent the murder of the victim, following his breach of protection orders.[39] Yet, it was of a different view in Kurt, where the applicants argued that pre-trial detention would have been an appropriate and proportionate measure in this case.[40] The Court extracted information from the GREVIO Baseline Report on Austria on the scarce use of pre-trial detention, even though this option is available under Austrian Law.[41] It found that despite pre-trial detention being an option, it had ‘no reason to call into question the authorities’ assessment that, on the basis of the information available to them at the relevant time, it did not appear likely that [the perpetrator] would obtain a firearm, go to his children’s school and take his own son’s life in such a rapid escalation of events’.[42] In making this statement the Court emphasised that Article 5 does not permit detention ‘unless it is in compliance with domestic law’,[43] yet while both Austrian and Bulgarian legislations allowed for it, the Court’s assessment of the assessment of risk seems uneven.  

22. The Court’s reasoning in relation to the (lack of) measures adopted by Austria in Kurt seems at odds with its own elaboration on the need to adopt a context sensitive assessment of the risk, taking into account the specific nature of domestic violence. Despite acknowledging that all the relevant factors need to be considered in domestic violence cases to properly assess the risk for all victims and the need for different measures, and having the input of GREVIO, the Court failed to recognise that the risk assessment in this case had been insufficient. It appears that, having established the ‘basic ingredients’ of a proper response to domestic violence cases (establishing a legal framework, an assessment the risk of violence, and providing effective protection) the Court more easily recognises violations to the Convention when State Parties ‘lack’ these ingredients, they are evidently insufficient (such as the absence of an offence of domestic violence, risk assessment tools or protection orders), or authorities persistently neglect to put them to use. The case of Kurt suggests, by contrast, that the Court still struggles with finding a violation where the State’s failure is not evident but is only revealed taking into account ‘the particular context of domestic violence’. Such a context, I argue, is not one determined by ‘factors’ alone, but relates to the broader structural system of inequalities in which domestic violence take place.

Domestic violence as a form of discrimination

23. Starting with the landmark Opuz judgment of 2009, the Court has often, yet not always, recognised that violence against women can constitute discrimination under Article 14 ECHR. In domestic violence cases in which the Court has found a violation, it has relied on reports by international and local human-rights organisations, periodic reports by human rights bodies, and statistical data from different authorities and academic institutions to build a sophisticated understanding of the drivers of domestic violence. Some of the key findings flowing from this input include the fact that domestic violence generally affects women more than men and the attitudes of law enforcement to the reporting of such allegations creates a climate of fear around the issue.[44] Should the State fail to protect women against domestic violence (noting that the failure does not have to be intentional) is in breach of their right to equal protection before the law.[45] 

24. Three of the recent cases help delineate the Court’s view of domestic violence as discrimination. In Tkhelidze, the Court has made significant strides towards a more comprehensive understanding of the complexity of gender-based violence. It carried out a ‘simultaneous dual examination’ of Articles 2 and 14, which allowed the Court to make a ‘contextual’ assessment of the different obligations of the State. As a result, the Court was able to effectively connect the inactivity of the authorities with the underlying gender-based discrimination and bias that provided fertile ground for the violence.[46] Moreover, the Court considered that persistent failure to address the issues identified, when assessed against the similar findings of other monitoring bodies, could be considered a ‘systemic failure’.[47] In Tunikova, the Court took another significant step forward by recognising that the persistent and systematic failure to address cases of violence against women by Russian authorities constitutes a failure ‘to create conditions for substantive gender equality that would enable women to live free from [violence] and to benefit from the equal protection of the law’.[48]

25. Another important development in these cases relates to the Court’s recommendations following the determination of the context of discrimination and gender bias and the consequent violation of Article 14. Firstly, the State must amend the consequences of such discriminatory treatment. For instance, when there is a suspicion of the possible role of gender-based discrimination in the commission of the crime, the State has ‘a pressing need to investigate the response of law enforcement’.[49] Secondly, the Court suggests that the State Party adopt different measures to counter or transform the context of discrimination. Examples of such measures are the implementation of an action plan to change the public perception of gender-based violence against women (as requested in other instruments); disseminating information on available remedies for victims; providing mandatory training on domestic violence dynamics for all personnel come into contact with victims (including judges); and establishing a monitoring mechanism and a way of recording the wealth of data on domestic violence disaggregated by sex, age and the relationship between perpetrators and victims.[50]

26. Despite these positive steps, as mentioned before, the Court continues to struggle to broaden its attention to inequalities and discrimination based on categories besides or in addition to gender when it comes to gender-based violence. This impacts on the Court’s understanding of the ‘particular context of domestic violence’, and consequently, its assessment of the risk and the appropriateness and proportionality of the measures adopted by the state. For instance, it seldom acknowledges the relevance of the migrant or racialised background of both victims and perpetrators and how these factors influence their social position and vulnerability.

27. This shortcoming is evident in Kurt, where the migrant background and associated socio-economic status of the applicant and the abuser were not considered. Judge Elósegui pointed to these shortcomings in her dissenting opinion in the latter case, stating that no information was given in the decision about the migrant background, educational attainment or socio-economic status of victim and perpetrator. The judge then highlighted that ‘according to the information provided by the applicant’s counsel during the hearing, the applicant was born in Turkey in 1978 and had her schooling there until she was fourteen years old. She moved from Turkey to Austria when she was fourteen. She attended school in Austria for only two years (from the age of fourteen to sixteen) in a low-level middle school (Hauptschule), where she started to learn German. She did not finish school and did not have any formal or professional training thereafter. She never attended a German course. She worked as a childminder and learned German from the children she was minding, as well as from her own children later on. Later she worked as a helper in a kitchen, and eventually lost her job. She was offered counselling in Turkish by a Turkish woman at the Centre for Protection from Violence because her German was not very good. When she went to the police she was accompanied by the Turkish woman from the Centre, but the interview with the police was in German and she had no interpreter’.[51] These elements speak of the victim’s social positioning, her dependency and increased vulnerability. It is striking that, despite making reference to documents and reports by different bodies, such as CEDAW or GREVIO, which have established the connection between an individual’s migrant status and their particular vulnerability to unequal treatment, including gender-based violence, the (majority of the) Court continues to overlook its relevance.

28. The lack of attention given to the structural factors affecting the overall social positioning of victims and perpetrators, despite continued calls for this from an array of third parties, restricts the ability of the Court to more fully understand the context within which violence takes place, particularly the link between inequalities of all kinds and the underreporting of/the factors that lead to the actual commission of violence. These factors ought to be considered also in relation to assessing the relevance of different factors in the process of risk assessment and the appropriateness of the protection measures adopted by the authorities. It appears that a comprehensive understanding of ‘the particular context of domestic violence’ will continue to elude the Court until it succeeds in understanding how intersecting inequalities contribute to gender-based violence.

L.P.A. Sosa

[1] Kurt v. Austria, ECtHR 4 July 2019, no. 62903/15, ECLI:CE:ECHR:2021:0615JUD006290315.

[2] Tkhelidze v. Georgia, ECtHR 8 July 2021, no. 33056/17, ECLI:CE:ECHR:2021:0708JUD003305617.

[3] Tunikova and Others t. Russia, ECtHR 14 December 2021, no. 55974/16, ECLI:CE:ECHR:2021:1214JUD005597416.

[4] A and B v. Georgia, ECtHR 10 February 2022, no. 73975/16, ECLI:CE:ECHR:2022:0210JUD007397516.

[5]  Y and Others v. Bulgaria, 22 March 2022, no. 9077/18 ECLI:CE:ECHR:2022:0322JUD000907718.

[6] Landi v. Italy, ECtHR 7 April 2022, no 10929/19,

[7] Opuz v. Turkey, ECtHR 9 June 2009, no. 33401/02, ECLI:CE:ECHR:2009:0609JUD003340102.

[8] Volodina v. Russia (no. 2), ECtHR 14 September 2021, no. 40419/19, ECLI:CE:ECHR:2021:0914JUD004041919.

[9] Talpis v. Italy, ECtHR 2 March 2017, no. 41237/14, ECLI:CE:ECHR:2017:0302JUD004123714.

[10] Kurt, para. 165; Tkhelidze, para. 48; Tunikova, para. 78.

[11] Tunikova, para. 86.

[12] Tunikova, para. 94.

[13] Kurt, para. 154.

[14] Kurt, para. 179; Tunikova, para. 95.

[15] Kurt, para. 178.

[16] Tunikova, para. 96.

[17] Directive 2012/29/EU of the European Parliament and the Council of 25 October 2012.

[18] Istanbul Convention, Articles 52 and 53.

[19] Tunikova, para. 157.

[20] Tunikova, para. 157.

[21] Kurt, paras. 177-83; Tunikova, para. 104.

[22] Kurt, paras. 179 and 181.

[23] Tunikova, para. 99.

[24] Kurt, para. 185.

[25] Kurt, para. 187.

[26] Kurt, paras. 165-6 and 175-6; Tkhelidze, para. 49; Tunikova, para 104.

[27] Tkhelidze, para 53.

[28] Kurt, para. 164.

[29] Kurt, para. 173.

[30] Tunikova, para. 104.

[31] Tunikova, para. 105.

[32] Kurt, para. 179; Tunikova, para. 95; Tkhelidze, para. 55.

[33] For a full comparative mapping of protection orders in Europe, see van der Aa S., Niemi J., Sosa L., Ferreira A., Baldry A. (2015) ‘Mapping the legislation and assessing the impact of protection orders in the European Member States’. For an analysis on emergency barring orders in particular, see Römkens, R. and Sosa, L. (2011) ‘Protection, Prevention and Empowerment: Emergency Barring Intervention for Victims of Intimate Partner Violence’, in Kelly, Hagemann-White, Meysen and Römkens (eds), Realising Rights. Case studies on state responses to violence against women and children in Europe, DAPHNE Project European Commission, (London), and Niemi, J. and Logar, R. (2017) ‘Emergency Barring Orders in Situations of Domestic Violence: Article 52 of the Istanbul Convention’. On the implementation of the EPO, see Cerrato, E. and others, ‘European Protection Order’ (2017). For a general discussion of the different aspects of protection order, see Sosa, Niemi and van der Aa, S. (2019); Protection against violence: the challenges of incorporating human rights’ standards to procedural law’, Human Rights Quarterly 41, No. 4: 939-961.

[34] See GREVIO Baseline Report, para 176.

[35] Austrian legislation was later amended to automatically include schools and formally inform school authorities of the existence of protection orders.

[36] Kurt, para. 165.

[37] Kurt, para. 183, see also Tunikova, para. 103.

[38] Kurt, para. 211.

[39] Y and Others, para 107 (b).

[40] Kurt, paras. 102 and 118.

[41] Kurt, para. 88.

[42] Kurt, para. 207.

[43] Kurt, para. 210.

[44] Opuz v. Turkey, ECtHR 9 June 2009, no. 33401/02, ECLI:CE:ECHR:2009:0609JUD003340102, paras.192-8; Tkhelidze, para. 56.

[45] Tkhelidze, para. 51.

[46] Tkhelidze, para. 60.

[47] Tkhelidze, para. 57.

[48] Tunikova, para. 129.

[49] Tkhelidze, para. 60; A and B, para. 44.

[50] Tkhelidze, para. 158.

[51] Kurt, dissenting opinion Judge Elósegui, para. 8.