1. Introduction
Police ill-treatment of racial and ethnic minorities, such as the Roma community, is a pressing issue in society today.1 Law enforcement agents, as representatives of the state, have human rights obligations to protect persons within the state’s jurisdiction from violence and not to discriminate against them. Even though police officers have a legitimate monopoly on the use of force, it is imperative that their actions be legal, necessary and proportionate – even when they are called on to make split-second decisions in tense, high-pressure situations.2 The literature abounds with materials on proper police training,3 anti-discrimination measures within policing,4 and domestic oversight mechanisms aimed at preventing the abuse of power and racist police violence.5 This article approaches the issue from a different angle, however. Researchers have found that ‘discriminatory practices by the police are not the result of individual pathologies but of perceptions of people that are shared within cultures and subcultures’.6 In other words, incidents of racism in policing tend to transcend the extent to which individual police officers share racist reviews. Rather, they are grounded in institutionalised practices which perpetuate social racism.7 These attitudes may undermine the efficacy of any domestic investigation into allegations of anti-Roma police violence (ARPV), underlining the importance of external oversight as provided by the European Court of Human Rights (ECtHR, the Court). Recognising the human rights implications of racist police violence, the article focuses on the role of the Court as a mechanism to ensure legal accountability for such abuses, in a context where domestic efforts to hold perpetrators to account may often be futile due to the very institutional biases identified above as a product of certain attitudes towards a community that are deemed acceptable by society as a whole.
This article examines and critiques the Court’s current approach in ARPV cases. These cases raise issues under Articles 2 (right to life), 3 (prohibition of ill-treatment) and 14 (prohibition of discrimination) of the European Convention on Human Rights (ECHR, the Convention). In cases concerning allegations of violence by state officials, the Court routinely distinguishes between the procedural limb and substantive limb of Articles 2 and 3 and separately assesses whether either was violated.8 In the ARPV cases examined, we see the Court extending this ‘bifurcation approach’ to its Article 14 assessment. This article examines the implications of this practice in discriminatory police violence cases.9 While the authors are aware that Roma are not the only group who may face abuse by law enforcement officers, and moreover that a bifurcation approach has been applied more generally, for instance, in cases concerning homophobic violence,10 the analysis has been limited to cases on ARPV. This specific focus enabled us to curate a sample which is sufficiently comprehensive to reveal patterns in cases of discriminatory violence committed by or with the acquiescence of state authorities, while remaining narrow enough to be subject to analysis. We have chosen to concentrate on anti-Roma violence specifically as a significant number of discriminatory police violence cases before the Court involve violence against members of this community.11
Following the brief introduction above, in Section 2 we explain how the bifurcation approach can be associated with a broader ‘procedural turn’ in the Court’s jurisprudence.12 Analysing relevant case law in Section 3, we explore how this approach has been applied in ARPV cases. While acknowledging the advantages of bifurcation in the context of Articles 2 and 3 of the Convention in Section 4, we submit that the particularities of the anti-discrimination clause and the issue of discrimination itself mean that a more integrated approach should be adopted in this area. We reflect on this finding in Section 5 and conclude in Section 6.
2. A procedural approach to substantive rights: quo vadis?
Complaints arising from instances of ARPV usually concern Articles 2 or 3 of the Convention as well as Article 14 invoked in conjunction with the former provisions. In this Section, we will explain how a procedural obligation to investigate has been read into the protective scope of Articles 2 and 3 leading to the adoption of a bifurcation approach. We explain how this is extended to the Court’s Article 14 assessment in APRV cases and how it can be associated with a broader procedural turn in the Court’s jurisprudence. We will touch on the arguments of subsidiarity and process efficacy associated with this development. We also highlight that additional procedural obligations can serve to ‘add teeth’ to the Court’s review. A more extensive discussion of the obligation to investigate, as applied in the case law, is reserved for Section 4.
2.1 Positive procedural obligations under Articles 2 and 3
Positive obligations are those ‘whereby a State must take action to secure human rights’.13 The Court has inferred positive obligations into Articles 2 and 3 although they are not immediately apparent from the text of the Convention itself.14 According to the Court’s own interpretation, these positive obligations are both procedural and substantive in nature.15 For instance, in X and Others v Bulgaria, the Court has explained that, in addition to the negative requirement to refrain from inflicting serious harm on persons within their jurisdiction,16 Article 3 implies further positive substantive obligations to (a) put in place a legislative and regulatory framework of protection and (b) take operational measures to protect specific individuals against a risk of treatment contrary to that provision in certain well-defined circumstances. It also implies a procedural obligation to effectively investigate any alleged violation thereof.17 Similarly, Article 2 implies a substantive obligation to put in place an appropriate legal and administrative framework defining when law enforcement officials may use force and firearms, as well as a procedural obligation to investigate any alleged violations.18
2.2 The procedural turn and positive procedural obligations
The reading of positive procedural obligations into substantive Convention rights as seen in the context of the ‘obligation to investigate’, has been associated with a broader ‘procedural turn’ in the Court’s jurisprudence.19 Various overlapping understandings of the precise contours of the procedural turn and the areas of the case law which fall within it can be identified in the literature.20 Janneke Gerards has created a useful two-pronged typology which will serve as the starting point for our analysis. Gerards’ typology distinguishes between the Court’s practice of (a) reading additional self-standing procedural obligations into the scope of substantive Convention rights and (b) cases where it has regard for procedural elements in its assessment of the proportionality of an interference.21 Current ECtHR judge, Oddný Mjöll Arnardóttir, has referred to the first limb of Gerards’ typology as the ‘procedural rights approach’ and the latter as ‘procedural review stricto sensu’.22 Procedural review stricto sensu is often applied in cases concerning qualified rights where a balance must be struck between competing rights or interests or where the contracting state has a wide margin of appreciation.23 Given the non-qualified nature of Articles 2 and 3 of the Convention and the issues which arise in the ARPV cases, it is the procedural rights approach which is relevant to our discussion.
While procedural obligations are explicit in Articles 5, 6 and 13 of the Convention, the procedural rights approach allows the Court to read them into other Convention provisions. They then become part of the protective scope of the right in question – as seen with regard to the duty to investigate in the cases of McCann v the United Kingdom24 (Article 2) and Assenov and Others v Bulgaria25 (Article 3) which we will analyse in detail below. Suffice to say for now that once procedural obligations have been established, the Court may choose to engage on the procedural or the substantive merits of the case – or indeed, on both.26 Positive procedural obligations can be read into any provision, but the approach is most common in the case law on Articles 2, 3, 8 (respect for private and family life), 10 (freedom of speech) and Article 1 Protocol 1 (protection of property)27.
2.3 The bifurcation approach
What is notable about its application in the context of the obligation to investigate, is that the Court does not just read additional positive procedural obligations into the scope of the right in question but adopts what can be termed a ‘bifurcation’ approach. This means that it distinguishes between the procedural limb and substantive limb of the provision and separately assesses if either of them has been violated. Thus, rather than contributing to an ‘overall’ violation of the Convention provision, procedural failings can lead the Court to find a partial, explicitly procedural violation. While this can also be seen in isolated cases under other Convention provisions, notably Article 8, when the right to an effective investigation becomes relevant,28 its ubiquity in the context of Articles 2 and 3 is striking.
Also striking is its extension to Article 14 as seen in the ARPV case law discussed further in Sections 3 and 4. The Court has frequently underlined that Article 14 is an ancillary provision, meaning that it only complements the other ‘core’ provisions of the Convention.29 In other words, it does not prohibit discrimination as such but only discrimination in enjoyment of the rights and freedoms set out in the Convention.30 Thus, Article 14 has ‘no independent existence’ and is always invoked in conjunction with another Convention provision.31 When assessing an Article 14 complaint in connection with a ‘core’ provision that has been ‘bifurcated,’ the Court often continues its ‘bifurcation’ approach – meaning in the context of Articles 2 and 3, that it separately assesses whether there was a discriminatory element in the substantive act of violence alleged and/or in the purportedly flawed investigation thereof.32 We argue that this extension of the bifurcation approach to Article 14 is not a self-explanatory practice. It is important to note that the finding of a violation of Article 14 in connection with the procedural limb of the non-ancillary Article has no impact on whether a violation of Article 14 in connection with the substantive limb is also found. Furthermore, the application of Article 14 does not necessarily presuppose a violation of the Convention right with which it is invoked in conjunction.33
2.4 The rationale behind the procedural turn
Seen in a broader context, the ‘procedural turn’ – as understood by Gerards,34 Arnardóttir35 and Kleinlein36 to encompass both the procedural rights approach and procedural review stricto sensu – has been characterised both as a response to the caseload crisis which threatened to overwhelm the Court subsequent to its judicial and geographical expansion during the 1990s,37 as well as evidence of the emergence of a new ‘age of subsidiarity’.38 In the context of the Convention, subsidiarity refers to the idea that the primary responsibility for securing the rights and freedoms defined in the Convention rests with the states themselves. While this principle has long been apparent in the jurisprudence of the Court,39 it was officially incorporated into the preamble of the Convention through the adoption of Protocol 15.40 The impetus for this can largely be attributed to the push for greater subsidiarity from traditionally Convention-compliant member states which felt that the Court was overstepping its role within the Convention system.41
The link between the procedural turn and subsidiarity is very clear when we consider procedural review stricto sensu and the deference associated therewith, as the Court displays a willingness to accept the outcome of domestic processes which comply with Convention requirements.42 In addition to institutional, subsidiarity-based arguments, process-efficacy rationale can be invoked to justify this deference. Process-efficacy rationale posits that good processes lead to good outcomes.43 In the context of the Convention, this can be understood to mean that convention-compliant processes are likely, although not guaranteed, to lead to convention-compliant outcomes.44 Although it may be less immediately obvious, arguments of subsidiarity and process efficacy can also be connected to the procedural rights approach as applied to the right to life and the prohibition of ill-treatment. The obligation to investigate any alleged violation of the Convention under Articles 2 and 3 can be seen as an effort to ensure that the responsibility to safeguard human rights remains primarily with the state itself. Furthermore, the procedural limbs of Articles 2 and 3 imply an obligation not just to investigate any alleged violations thereof but to do so properly – in a manner that complies with the specific procedural standards set by the Court.45 Although the requirement to investigate is an obligation of means not results,46 process-efficacy rationale suggests that through meticulous oversight of procedural integrity, the Court can ensure its effectiveness.
A perhaps more obvious justification for the procedural rights approach is its ability to ‘give teeth’ to the Court’s review, particularly in cases where states have a broad margin of appreciation or where a substantive violation of a particular Convention right is difficult to prove.47 For example, in many Article 2 cases concerning extrajudicial killings, it may be impossible for the Court to establish the involvement of a state agent ‘beyond reasonable doubt’, as the precise circumstances in which the violence occurred are only known to the state.48 Despite this, however, a violation may still be found if the state is shown to have failed in its procedural obligation to adequately investigate the circumstances of the applicant’s death.49 In the next Section, we will explain how this approach can be seen in ARPV cases before reflecting on the potential consequences thereof.
3. Exploring the case law: Articles 2, 3 and 14 in ARPV cases
In this Section we first provide an overview of the cases which we examined and the conclusions reached by the Court. Only then will we move to discuss the cases in more depth.
3.1 Overview of the cases examined
Our corpus consists of 52 cases concerning ARPV.50 This sample includes all cases identified involving police violence against Roma people which led to a judgment on the merits. Note, that this means that our sample also includes cases where (a) a discrimination claim was raised by the applicants but declared inadmissible or not separately examined (however, the Court did assess the merits of the claim under the core Convention provision), as well as (b) cases against Roma where no Article 14 claim was raised.51 We included this latter category of cases as interviews with practitioners have revealed that discrimination claims may not be raised due to their low success rate or other strategic considerations.52 As this paper focuses on the application of a bifurcation approach to Article 14, neither of these two categories will be subject to great analysis. They are, however, included in Figure 1 (below) and in this case assessment to provide a comprehensive overview of the Court’s jurisprudence on ARPV.
Figure 1
To compile the corpus, a search of the Court’s online database (HUDOC) was carried out in respect of Articles 2 and 3 alone and together with Article 14. The scope of the search was limited by excluding keywords related to the core articles which are not relevant to our current focus.53 Additional search terms of ‘police’ and ‘Roma’ were added in English, and ‘l’ethnie rom’, ‘d’origine rom’, ‘groupe ethnique des Roms’ and ‘gens du voyage’ in French. In a final step, cases in which the word ‘Rom*’ appeared in a geographical context or which concerned violence committed by private parties, were manually removed. The latter limitation reflects our focus on ARPV as it concerns a direct abuse of power by state agents (substantive aspect) on the one hand, and often indicates the presence of institutional bias (procedural aspect) against Roma on the other. This search method, complemented by adding cases from literature, or at the suggestion of practitioners, yielded 53 results up to and including February 2023.
The table in Figure 1 provides a breakdown of ARPV cases. In this table, each block represents one case. To provide an insight into the types of violence experienced in these cases, we have divided the cases into those concerning violations which occurred while the victim was in or outside police custody. The left column contains all of the cases which occurred outside of custody, mostly in public spaces,54 but some in the homes of Roma,55 during raids and searches,56 or as part of pogroms57 or extrajudicial executions.58 The middle column concerns arrests (for example by the use of force) or abuse that occurred during transport.59 The right column shows cases which occurred in custody. These are typically forced confessions60 but there were also four cases where the victim inexplicably fell out of a police station window.61
As seen in the key, the outline surrounding the block represents the type of core right violation found by the Court. In 32 cases (62%), the Court found a violation of both the substantive limb and the procedural limb of Articles 2 and/or 3, the core Article(s), raised. In two cases, the Court found only a substantive violation of the core right. In these cases, the procedural shortcomings were so connected to the substantive limb, or so severe that the Court did not find it necessary to separately examine both limbs,62 but considered them together.63 In 12 cases (23%) it found only a procedural violation of the core Article, meaning that responsibility for the abuse itself was not or could not be established; however, the quality of the domestic investigative proceedings was so inadequate that they could never have established who was responsible. In one case, the claim relating to the core right was rejected.64
The outcomes of the complaint under Article 14 are represented through the colours of the blocks. In comparison to the 32 cases where the Court found both a procedural and substantive violation of the core provisions invoked, there were only two cases (4%) where both a substantive and procedural violation of Article 14 were found.65 Even excluding the ten cases where no Article 14 claim was raised by the applicant,66 this is a significant discrepancy. In 13 cases (25%), the Court found a procedural violation of Article 14, meaning that the Court was dissatisfied with the level of scrutiny or attention given in the investigation to unmasking possible racist undertones – potentially due to biased attitudes of the investigative authorities as noted by the Court itself in certain cases.67
It should be noted that in ten cases where only a procedural violation of Article 14 was found, both a procedural and a substantive violation of the core provision had been established. In other words, in these ten cases (23%), despite establishing that the violence had been the responsibility of the State authorities and that the ensuing investigations were unsatisfactory (either because potentially racist elements of the abuse had not been properly investigated, or because the inadequacy of the procedures or other procedural shortcomings were the result of bias),68 the Court did not find that it had been proven beyond reasonable doubt that the violence itself was driven by racist motives. In eight cases (15%), the Court found that the procedural limb of Article 14 had not been violated, mostly because it deemed the evidence not sufficient to have warranted a separate line of inquiry in the domestic proceedings.69
As an interim conclusion before offering our commentary on the cases, we can observe that the above initial comments on the cases echo those made by other scholars about cases concerning racist violence more generally.70
3.2 Commentary on the cases
Arguably, the discriminatory element of police violence is often elusive from a legal evidentiary perspective, which is (at least partly) the reason why the Court finds fewer violations of Article 14 than Articles 2 and 3, both in absolute numbers and proportionately. For some commentators, the discrepancy found in the number of procedural and substantive violations indicates a reluctance on the part of the Court to find a state responsible for racism.71 While it is difficult to conclusively prove such an allegation, it is clear that criticising a state for its failure to properly investigate an instance of racist police violence is less serious than finding that its agents directly engaged in racist violence. While a hostile climate against Roma is well documented in several member states, it is incredibly difficult to prove the impact of structural issues and systemic biases in individual cases.72 Other critics of the ARPV case law specifically, and the Court’s evidentiary considerations more generally, have written about the difficulties which the Court’s requirement of proof beyond reasonable doubt means in practice.73
Under the Court’s current bifurcation approach to Article 14, it first assesses whether there was a discriminatory intent behind the specific act of violence itself. Provided there are grounds for a reasonable suspicion that there was (even if this has not been proven), in addition to the pre-existing obligation to conduct an investigation into whether the violence took place, the Court must assess whether the allegedly discriminatory element thereof was properly investigated.74 So far explicit racial profiling and the use of racial slurs have been viewed by the Court as such indicators,75 while stereotypes,76 or derogatory words were not in themselves sufficient to warrant a specific inquiry into the discriminatory element.77 In other words, to find a substantive violation of Article 14 the Court requires direct evidence that the specific act of violence was accompanied by discriminatory intent. It does not recognise the evidentiary value of proof of a climate of discrimination in itself or even as justification to shift the burden of proof onto the State to show that racism did not play a role in the case at hand.78
When contrasting the Court’s case law on violent with non-violent instances of discrimination against Roma,79 – for example, judgments concerning school segregation cases – or violent forms of discrimination against Roma with other forms of discriminatory violence,80 – for example, judgments concerning domestic violence cases – the difference in the Court’s approach is striking.81 In both these other groups of cases, the Court has issued some ground-breaking judgments where it was more receptive to proof of systemic issues. In the education segregation case of DH and Others v the Czech Republic, the Grand Chamber accepted statistical data as sufficient evidence about the discriminatory placement of Roma children in special schools.82 In the domestic violence case of Opuz v Turkey the Chamber accepted reports that the negligence of domestic authorities constituted a form of discrimination against women.83 In relation to the disconnect between these lines of case law (namely, racial discrimination with or without violence, or violence based on different discrimination grounds) much has been written about the applicable standard and burden of proof,84 the types of evidence worth considering,85 and even the role of the judges’ subconscious,86 to try and explain what is preventing the Court from holding in more cases that the abuse which Roma and other racialized minorities suffer at the hands of police occurred in a racist context.
There seems to be an agreement among scholars that the very high standard of proof – beyond reasonable doubt – required by the Court in racist violence cases is problematic. Mačkić argues that the Court’s approach to indirect discrimination should be applicable to racist police violence cases, as opposed to the current direct discrimination approach, thereby expanding the types of evidence on which the applicants may rely.87 Henrard highlights the lack of clarity on the level of persuasion both at prima facie and full proof stages, and calls for an acknowledgment when the burden of proof should shift from the applicant to the respondent state.88 Dembour, Rubio-Marín and Möschel, and Havelková have explored historical and political reasons which may help to explain some of the judges’ hesitant approaches to anti-discrimination law in general and to the classification of violence as a form of discrimination in particular.89 Henrard also argues that, for reasons of legitimacy, the Court tends to concern itself with not imposing too heavy a burden on the state.90 These theories do not provide a decisive explanation of why the Court has adopted such an approach. They do, however, highlight a general consensus in the literature that the Court’s current approach is problematic and renders it particularly difficult for applicants to ensure that the state is held to account.
3.3 Conclusions from the case law
When faced with the issue of discrimination in ARPV cases, this article asks whether the Court’s approach of separately assessing the presence of discrimination in relation to the substantive and procedural elements of the violation of the core provision is a just one; meaning is it an appropriate way to ensure that the applicant’s rights are properly protected without placing an unfair burden on the state? It is self-evident that any sort of recognition of the discriminatory element in these cases is preferable to a failure to acknowledge it entirely. The Court’s current approach is also preferable to one whereby it rejects the discrimination claim altogether or refrains from examining Article 14 separately, simply by merging the discrimination claim into its assessment of the core Article.91 However, it is apparent from our review of the case law, that the bifurcation model is neither ideal nor the only avenue available to the Court. In the remainder of this article, we endeavour to illustrate that a ‘global’ approach to Article 14 would be preferable. Under such an approach, the Court would jointly assess whether there was a discriminatory element in both the commission of violence and the investigation into it to find an ‘overall’ violation of Article 14. Such an approach has been advocated for by the European Roma Rights Centre (ERRC)92 and could provide more evidentiary flexibility, thereby allowing the Court to ‘remarry’ the procedural and substantive when it comes to Article 14. Evidence of discrimination in relation to the investigation or the act of violence in itself would contribute equally to proving the state’s responsibility. Having given a brief overview of the issues raised and the types of violations found in ARPV cases, let us now take a closer look at the approach taken by the Court and the merits and drawbacks thereof.
4. Positive procedural obligations in ARPV cases: the duty to investigate
In this Section, we consider the duty to investigate under Articles 2, 3 and 14 in more detail. In doing so, we examine the origins of the duty to investigate in the context of these three Articles, and analyse the case law on ARPV to understand how these principles have been applied in these cases.
4.1 The duty to investigate under Articles 2 and 3
In the context of Article 2, the obligation to investigate the circumstances of deaths occurring under state control can be traced back to the Court’s ruling in McCann and Others v the United Kingdom, where it found that the general obligation under Article 1 to secure Convention rights to everyone, required by implication that there be some form of effective official investigation when individuals have been killed as a result of force by agents of the state.93 Since then, the Court has accepted that the obligation arises in a variety of situations where the individual has sustained life-threatening injuries, died, or disappeared in violent or suspicious circumstances.94 The Court employed the same logic to infer that an identical obligation to investigate exists under Article 3 in Assenov and Others v Bulgaria,95 which was, incidentally, the first ARPV case which the Court ever considered. In Labita v Italy the Court further explained that, in the absence of an obligation to investigate, Article 3 would be ineffective in practice and, in some cases, it would be possible for state agents to abuse the rights of those within their control with virtual impunity.96 Furthermore, a proper response by the authorities to alleged violations of Article 3 is essential to maintain public confidence and to prevent any appearance of collusion or tolerance of unlawful acts.97 The obligation to investigate has become a firmly established element of both of these provisions.
As mentioned in Section 3.1, there were no Article 2 cases in the sample where the Court found neither a substantive nor a procedural violation.98 In 71% of the ARPV cases concerning Article 2, the Court found a substantive violation, holding a contracting state responsible for the life-threatening injury or the loss of life occasioned by the police violence. In the few cases where there was not sufficient evidence to establish that the authorities were directly responsible for the life-threatening situation, or loss of life, the respondent states were at least condemned for their failure to investigate the allegations – thus, in 90% of the cases a procedural violation was found.99 Let us focus on the cases where in the absence of a substantive violation the Court could still establish a procedural one. These instances, we submit, are the ones where the procedural assessment shows its ‘applicant-friendly’ side by highlighting that without the additional layer of protection required by the duty to effectively investigate allegations of ill-treatment or death in custody, abuse of power could go unpunished.
The Court’s reasoning is very telling in these judgments. It generally begins by acknowledging its subsidiary role, refraining from reassessing the facts as established by domestic authorities. With regard to the substantive aspect of Article 2, its reason for not finding a violation is not that none took place per se but that ‘there is no evidence available to the Court’100 or that ‘there is insufficient factual and evidentiary basis on which to conclude’101 that a violation of the substantive aspect of Article 2 took place. Generally, it is only after establishing that there is insufficient evidence of a substantive violation that the Court assesses the theoretical effectiveness of the domestic framework,102 or indeed whether the investigative proceedings in practice have met the standards set by the Convention. This is not a hard rule, however, as in the case of Fedorchenko and Lozenko v Ukraine for example, the Court began by considering the efficacy of the investigation and finding a procedural violation of Article 2.103 It then held that ‘in the absence of other evidence, and given the above conclusion that there was no effective investigation in the present case, the Court cannot draw a conclusion beyond reasonable doubt’104 that a substantive violation took place. By examining the implied positive duty to conduct effective investigations into alleged deaths and finding self-standing violations thereof, the Court prevented domestic authorities from evading any sort of accountability for serious cases of abuse of power.
The same pattern can be seen in the case law on Article 3. In 60% of the cases the Court found a substantive violation and in 80% a procedural violation. Similar to Article 2, under Article 3 a bifurcation approach could already be seen in the previously mentioned 1998 case of Assenov and Others v Bulgaria, relating to alleged police violence against a 14-year-old Roma boy who sustained injuries while in police custody. It was not disputed that the boy had been the victim of violence, but there was insufficient evidence to establish that it had been perpetrated by the police.105 The evidence was sufficient, however, to establish a ‘reasonable suspicion’ that the injuries may have been caused by the authorities, generating a positive obligation for the state to effectively investigate the circumstances thereof. The inefficiency of the investigation allowed the Court to find a procedural violation of Article 3 even though no substantive violation could be proved.106
In a number of judgments where only a procedural violation of Article 3 was found (26%),107 the Court highlighted that its inability to establish the facts that would allow it to find a substantive violation was owed ‘primarily to the national authorities’ inactivity and reluctance to carry out an effective investigation into the applicant’s allegations’.108 In other words, ‘this impossibility stems largely from the absence of a thorough and effective investigation of the applicant’s complaint’.109 In other cases, it noted that there was simply no investigation into the applicant’s claims,110 or that its subsidiary role prevented it from substituting its own conclusions based on disputed facts.111 In the remainder of the cases, the Court assessed whether the criteria for investigating allegations of ill-treatment in custody (later detailed and solidified in the case of Bouyid v Belgium) were met.112 This Grand Chamber decision established that for an investigation to be effective those responsible for carrying it out must be independent; the authorities must be proactive; the investigation must be capable of identifying and punishing those responsible; it must be conducted promptly and thoroughly and the victim must be able to participate effectively in the investigation.113 At the same time, it is interesting to note that, rather than contributing to a finding of an ‘overall’ violation or being seen as evidence contributing to a finding of such, the state’s failure to investigate is treated as a distinct type of qualified violation: a procedural one. This ‘bifurcation’ approach is consistently adopted regardless of whether Article 2 or Article 3 is invoked.114
4.2 The duty to investigate under Article 14
From a certain perspective it seems almost intuitive that when the procedural and substantive elements of Articles 2 and 3 are separated, the same approach would be applied by the Court to Article 14 in light of its ancillary nature. In this Section, we consider the emergence of this style of review in the context of the duty to investigate under Article 14. We highlight some key issues with its application and argue in favour of a more global and integrated assessment.
The Court’s current approach to Article 14 can be traced back to the partly dissenting opinion of Judge Bonello in Anguelova v Bulgaria where he raised it as a possible avenue to render the provision practical and effective. This was intended as a measure to rectify the existing situation – whereby no violation of Article 14 at all was ever found in the context of ARPV. In numerous cases, applicants attempted to rely on Article 14 but the Court rejected their complaints as unsubstantiated, or found in favour of the state.115 This was the backdrop against which Judge Bonello delivered his partly dissenting opinion. He criticised the Court for its inability to recognise that the violation of disadvantaged minority groups’ rights including Kurds, Muslims and Roma is not simply a result of ‘well-disposed coincidence’.116 He proposed the adoption of the type of bifurcation approach common under Articles 2 and 3 as a potential solution:
The Court has also, by an admirable process of judicial activism ‘created’ the concept of a ‘procedural violation’ (…). The self-same rationale that found in a non-investigation, or an inadequate investigation of death or inhuman treatment by the State, a ‘procedural violation’ of those guarantees, should inspire and would justify the finding of a violation of Article 14 taken in conjunction with Articles 2 or 3 where no proper investigation of the alleged violation has been carried out.117
Four years later, a distinction was drawn between the ‘procedural and substantive aspects of Article 14, taken together with Article 2’ by the First Section of the Court in the Chamber judgment in Nachova and Others v Bulgaria.118 This approach was subsequently confirmed in the Grand Chamber judgment in the same case in 2005.119 Since then, to the best of our knowledge, there have been no cases in this area where the Court has not taken a bifurcation approach to the assessment of an Article 14 claim. However, just because an approach has become established, it does not mean that it should be accepted. In fact, six judges voiced their discomfort with the majority’s approach to the Article 14 complaint in the Grand Chamber judgment in Nachova and Others v Bulgaria:
We cannot subscribe to the new approach adopted by the Court which entails linking a possible violation of Article 14 of the Convention to the substantive and procedural aspects of Article 2 individually. An overall approach would have been preferable, since it would have better reflected the special nature of Article 14, which has no independent existence (…) [therefore] we consider it artificial and unhelpful to distinguish between the substantive and procedural aspects, especially as in the instant case the Court found violations of both these aspects of Article 2 [emphasis added].
To reiterate, the results of our case law analysis revealed a striking disparity between the number of cases where the Court found both a procedural and substantive violation of Articles 2 or 3 (N = 32) as opposed to Article 14 (N = 2).120 In ten cases (23%)121 the Court found only a procedural violation of Article 14 despite having already found both a substantive and procedural violation of the core Article. This means that, despite establishing that the violence had been the responsibility of the State authorities and that the investigations did not meet Convention requirements, the Court did not find that it was shown beyond reasonable doubt that the violence, or the lack of its thorough investigation had been inspired specifically by racist motives. Arguably, it is this high standard of proof which necessitates the Court taking a procedural approach in the first place. This sets the bifurcation of Article 14 apart from the bifurcation of Articles 2 and 3 as in the latter context it serves to provide additional protection against a clearly deficient State.
As currently applied by the Court, a bifurcation approach to Article 14 puts most emphasis on whether the domestic authorities should have investigated any potential racist element of the complaint in light of specific indicators in the individual case.122 This approach is understandable in cases where there are also evidentiary hurdles to proving that a substantive violation of the core provision took place123 – how can we say violence was discriminatory when we cannot say for certain that it occurred? However, it becomes more difficult to accept in cases where a substantive violation of Articles 2 or 3 has already been established.124 Furthermore, as pointed out by the ERRC in many of the cases it supported, the separate assessment of the investigative procedures from a discrimination point of view is logically flawed: it is difficult to imagine that an investigation which the Court had already found to be ineffective could be effective in the aspect of uncovering racist elements of the incident.125 Indeed, the only cases involving both Articles 3 and 14, where the Court did not find both a procedural violation of first Article 3 and then Article 14 in conjunction with Article 3, are those where the Article 14 claims were deemed ill-founded.126 It is not difficult to see that the involvement of law enforcement agents raises additional questions and concerns about public confidence in the authorities’ adherence to the rule of law, and the appearance of collusion in, or tolerance of, unlawful acts, which makes it even more important to give the allegations due consideration.127
Lingering like smoke and tainting everything it touches, discrimination is a pervasive and elusive force. Notoriously difficult to prove, any overt manifestation thereof could be taken as evidence of a broader set of covert biases within an institution or society.128 We submit that any evidence of discrimination at the procedural level may indicate a larger institutional issue and therefore should not be viewed in a vacuum. Our conclusion is that the Court’s procedure-focused approach effectively disregards the evidentiary value of an overall context of a general discriminatory climate against racial and ethnic minorities. In the context of discrimination in particular, divorcing the procedural from the substantive creates a false dichotomy between two integrated elements of the same complaint and creates an arbitrary separation between them. Both the actual direct, racist ill-treatment of an individual as well as a reluctance or failure to investigate alleged instances thereof – again due to their ethnicity – often result from the self-same system of bias against racialized members of society. There may be two separate actions but, ultimately, they are both illustrations of the same underlying issue, discrimination, which is already very difficult to prove.
5. Reflecting on the analysis
It is crucial to note that Judge Bonello himself had other proposals for how the Court could recognize that the issue of potentially racist police abuse of vulnerable minorities is a complex and difficult one. He writes:
Ideally [the Court] should reconsider whether the standards of proof should not be the more juridically justifiable ones of preponderance of evidence or of a balance of probabilities. Alternatively, it should, in my view, hold that when a member of a disadvantaged minority group suffers harm in an environment where racial tensions are high and impunity of State offenders epidemic, the burden to prove that the event was not ethnically induced shifts to the Government. Subordinately, in the sphere of Article 14, as it has done in the case of Articles 2 and 3, the Court ought to invest in its own doctrine of ‘procedural violation’ (…). [Empasis added]129
Indeed, there have been several cases where government failure to submit information to which only it could have access gave rise to inferences that the charges against the state were well-founded.130 The application of such an approach to cases of ARPV would much more accurately reflect the integrated and systemic nature of discrimination. In the few cases where the Court did find both a substantive and procedural violation of Article 14 taken in conjunction with Article 3, it took the overall context of the authorities’ failure in their duty to unmask racial motives for the ill-treatment (procedural violation) and examined the implications of this finding on the substantive aspect of the case.131 Phrased differently, in these cases ‘the Court consider[ed] that the manner in which the authorities justified and executed the police raid show[ed] that the police had exercised their powers in a discriminatory manner’132 which led to the finding of both a substantive and procedural violation.
It is clear that there is a special stigma associated with the finding of a substantive violation of a Convention provision.133 In light of this, the particular nature of discrimination and how difficult it is to prove, instead of a rigid bifurcation approach when it comes to Article 14 the Court should conduct a more ‘global’134 assessment, asking ‘whether the incident [as a whole] took place in a context where the authorities, aware of problems of police targeting Roma for violence, allow police forces contaminated with institutional anti-gypsyism to act with impunity’.135 In examining whether domestic authorities complied with their obligations to investigate, the Court should acknowledge that, ‘particularly in cases of systemic discrimination, the flawed character of investigations into alleged discriminatory measures/actions can itself be a manifestation of discrimination. In other words, the procedural and the substantive dimensions of the prohibition of discrimination are connected’.136 As such the evidentiary value of procedural failings should not be limited to proving discrimination in the investigation but rather be recognised for what they (often) are – manifestations of a climate of discrimination and bias. In the ARPV case law such an approach would mean that, if discrimination was uncovered in either the commission of, or the investigation into, the police violence, it could lead to an unqualified violation of Article 14 in conjunction with the relevant core provision(s) as it would be taken as indicative of wider institutional biases. We propose further research into precisely this question: the potential of procedural failings as an indication of a substantive issue.
While accepting the efficacy of the bifurcation approach under Articles 2 and 3, we submit that (re-)marrying all elements of Article 14 into findings of global violations would better reflect the reality of discrimination and be preferable from the point of view of the individual applicant, and from the point of view of trust in public authorities and the rule of law. To the argument that the Court should not impose a too heavy burden on the respondent state we submit, in full agreement with Roberts, that the costs of erring on the side of caution that benefit the state far outweigh the ‘harms’ that the finding of a violation may carry. Firstly, because
[h]uman rights violations are rarely entirely isolated, moreover, such that failing to recognise meritorious rights claims will often allow one or another form of rights violation to continue. (…) [And secondly,] the cost of a state losing an international human rights judgment is reduced, insofar as orders for the state to comply with its rights obligations may be easily observed where the state was complying with such orders in any case.137
6. Conclusion
This article analysed the Court’s case law in incidents of ARPV. It found that the Court has taken the approach to separately examine substantive and procedural aspects of both the abuse (within the scope of Articles 2 or 3) and the discrimination element (under Article 14). This separation can be traced back to a more general ‘procedural turn’ in the Court’s jurisprudence – a key aspect of which involves the reading of positive procedural obligations into the scope of substantive Convention rights. Although the procedural turn is grounded in the principle of subsidiarity, it is also supported by arguments of process-efficacy (just processes lead to just outcomes) and the development of the additional obligation to conduct an effective investigation can help strengthen the Court’s review in cases where the precise facts cannot be established.
Analysing the relevant body of case law, consisting of over 50 judgments, we found that substantive violations were common in Article 2 and Article 3 cases taken alone, and an additional procedural violation was found in an even larger ratio of these cases. It is submitted that separately assessing the quality of investigations may enable the Court to ‘catch’ violations of core rights even where the evidence before it is insufficient to hold the respondent state responsible for the abuse itself. However, when it comes to Article 14, the number of substantive violations found by the Court is almost negligible, which poses the question of whether a procedural approach is capable of addressing the essential element of these cases: the generally hostile, anti-Roma climate which exists in some member states. We argued that in cases of discriminatory violence by law enforcement agents any flaws in the investigations might themselves be a form of racist discrimination, and therefore that substantive and procedural elements of Article 14 cannot be artificially separated.
Notes
[1] European Roma Rights Centre, Brutal and Bigoted: Policing Roma in the EU (2022) European Roma Rights Centre.
[2] S Casey-Maslen, ‘Academy in-brief no. 6: Use of Force in Law Enforcement and the Right to Life: The Role of the Human Rights Council’ (Geneva Academy of International Humanitarian Law and Human Rights, 2016), 6–9; Council of Europe, ‘The European Code of Police Ethics Recommendation Rec (2001)10 of the Committee of Ministers to member states on the European Code of Police Ethics’ (Adopted by the Committee of Ministers on 19 September 2001), paras 35–40; Bouyid v Belgium (GC), App no. 23380/0 [2015] ECtHR, para 100: ‘where an individual … is confronted with law-enforcement officers, any recourse to physical force which has not been made strictly necessary by the person’s conduct diminishes human dignity and is in principle an infringement of the right set forth in Article 3 of the Convention’.
[3] P Costa & I Thorens, Training Manual on Police Integrity (Geneva Center for the Democratic Control of Armed Forces, 2015); Council of Europe, Committee of Ministers, Toolkit: International Police Standards. The European Code of Police Ethics (Geneva Center for the Democratic Control of Armed Forces, 2001).
[4] European Union Agency for Fundamental Rights, Preventing unlawful profiling today and in the future: a guide (Publications Office of the European Union, 2018); OSCE High Commissioner on National Minorities Recommendations on Policing in Multi-Ethnic Societies (OSCE, 2006); R Oakley, Policing Racist Crime and Violence: a Comparative Analysis (European Monitoring Center on Racism and Xenophobia, 2005).
[5] I Scott & M Lisitsyna, ‘Who polices the police? The role of independent agencies in criminal investigations of state agents’ (2021), Open Society Justice Initiative, DOI:10.34880/74m3-9s14; OSCE International Police Standards: Guidebook on Democratic Policing (DCAF, 2009); United Nations Office on Drugs and Crime, Handbook on police accountability, oversight and integrity (United Nations, 2011).
[6] B Rafael, ‘Verdacht und Vorurteil. Die polizeiliche Konstruktion der „gefährlichen Fremden’ in C Howe & L Ostermeier (eds.), Polizei und Gesellschaft: Transdisziplinäre Perspektiven zu Methoden, Theorie und Empirie reflexiver Polizeiforschung (Springer VS Wiesbaden, 2019), 38.
[8] The procedural limb concerns the obligation to investigate while the substantive limb concerns the alleged act of violence itself.
[10] Aghdgomelashvili and Japaridze v Georgia, App no. 7224/11 [2020] ECtHR; MC and AC v Romania, App no. 12060/12 [2016] ECtHR. Note that in the latter case the violence was not committed by the police.
[11] S Latal et al., ‘Global Focus On Police Brutality Strikes Chord In Southeast Europe’ (2020) Balkan Insight, 12 June 2020, <https://balkaninsight.com/2020/06/12/global-focus-on-police-brutality-strikes-chord-in-southeast-europe/>.
[12] O Arnardóttir, ‘The “Procedural Turn” under the European Convention on Human Rights and Presumptions of Convention Compliance’ (2017) 15 International Journal of Constitutional Law, 1.
[13] D Harris et al., Law of the European Convention on Human Rights (Oxford University Press, 2009), 18.
[14] J Akandji-Kombe, Human Rights Handbook, No. 7: Positive Obligations under the European Convention on Human Rights: A Guide to the Implementation of the European Convention on Human Rights (Directorate General of Human Rights, Council of Europe, 2007). See further: V Stoyanova, Positive Obligations under the European Convention on Human Rights: Within and Beyond Boundaries (Oxford University Press, 2023).
[17] See discussion of the duty to investigate Article 3 in COE, Guide on Article 3 of the European Convention on Human Rights: Prohibition of Torture (COE, updated 31 August 2022), 26–34. See similarly in the context of Article 2 COE, Guide on Article 2 of the European Convention on Human Rights: Right to Life (COE, updated 31 August 2022), 32–46.
[18] Pietrzak and Bychawska-Siniarska and Others v Poland, App nos. 72038/17 and 25237/18 [2024] ECtHR, para 209.
[20] See e.g. E Brems, ‘Procedural Protection: An Examination of Procedural Safeguards Read into Substantive Convention Rights’ in E Brems & J Gerards (eds.); E Brems & J Gerards, Shaping Rights in the ECHR: The Role of the European Court of Human Rights in Determining the Scope of Human Rights (Cambridge University Press, 2013); B Çalı, ‘Coping with Crisis: Whither the Variable Geometry in the Jurisprudence of the European Court of Human Rights’ (2018) 35 Wisconsin International Law Journal, 237; O Arnardóttir, ‘Organised Retreat? The Move from “Substantive” to “Procedural” Review in the ECtHR’s Case Law on the Margin of Appreciation’ European Society of International Law (ESIL) 2015 Annual Conference (2015); Arnardóttir (n 12).
[21] J Gerards, ‘Procedural Review by the ECtHR—A Typology’ in J Gerards & E Brems (eds.) Procedural Review in European Fundamental Rights Cases (Cambridge University Press, 2017), 127–160.
[23] See for example: LB v Hungary (GC), App no. 36345/16 [2023] ECtHR; Axel Springer SE v Germany, App no. 8964/18 [2023] ECtHR; Halet v Luxembourg (GC), App no. 21884/18 [2023] ECtHR.
[26] See for example, Roşioru v Romania, App no. 37554/06 [2012] ECtHR; Suleymanov v Russia, App no. 32501/11 [2013] ECtHR; Tarasov v Ukraine, App no. 17416/03 [2013] ECtHR.
[28] See for instance, the Article 8 cases: Erkan Birol Kaya v Turkey, App no. 38331/06 [2018] ECtHR; Khadija Ismayilova v Azerbaijan, App nos. 65286/13 and 57270/14 [2019] ECtHR; EG v The Republic of Moldova, App no. 37882/13 [2015] ECtHR.
[29] The Court uses the term ‘substantive’ to refer to the other ‘non-ancillary’ Convention provisions which Article 14 may complement. We have chosen to use the term ‘core’ to avoid confusion due to the distinction already being drawn in this paper between ‘procedural’ and ‘substantive’ obligations.
[30] COE, Guide on Article 14 of the European Convention on Human Rights and on Article 1 of Protocol No. 12 to the Convention: Prohibition of Discrimination (COE, updated on 29 February 2021), 6.
[33] Carson and Others v the United Kingdom (GC), App no. 42184/05 [2010] ECtHR, para 63; EB v France (GC) (n 31).
[36] T Kleinlein, ‘The Procedural Approach of the European Court of Human Rights: Between Subsidiarity and Dynamic Evolution’ (2019) 68 International and Comparative Law Quarterly no. 1, DOI: 10.1017/S0020589318000416, 91–110.
[38] R Spano, ‘Universality or Diversity of Human Rights? Strasbourg in the Age of Subsidiarity’ (2014) 14 Human Rights Law Review no. 3, DOI: https://doi.org/10.1093/hrlr/ngu021, 487.
[39] J Laffranque, ‘Subsidiarity: From Roots to Its Essence’ (Speech at the seminar traditionally held to mark the opening of the judicial year of the European Court of Human Rights, 2015) <https://www.echr.coe.int/documents/d/echr/Speech_20150130_Seminar_Laffranque_2015_BIL> (last visited 8 September 2023); Case ‘Relating to Certain Aspects of the Laws on the Use of Languages in Education in Belgium’ v Belgium (Merits), App nos. 1474/62, 1677/62, 1769/63 and 2126/64 [1968] ECtHR; Handyside v the United Kingdom, App no. 5493/72 [1976] ECtHR, para 48.
[41] O Stiansen & E Voeten, ‘Backlash and Judicial Restraint: Evidence from the European Court of Human Rights’ (2020) 64 International Studies Quarterly no. 4, DOI: https://doi.org/10.1093/isq/sqaa047, 770.
[42] See for example, Avci v Denmark, App no. 40240/19 [2021] ECtHR, Von Hannover v Germany, App no. 59320/00 (ECtHR 24 June 2004).
[44] E Brems, ‘The “Logics” of Procedural-Type Review by the European Court of Human Rights’ in Gerards & Brems (n 21), 21.
[45] Note that, as mentioned previously, an obligation to investigate can also arise under other Convention provisions. See under Article 8 e.g. Erkan Birol Kaya v Turkey, App no. 38331/06 [2018] ECtHR.
[46] See, regarding Article 2 of the Convention, McCann and Others v the United Kingdom, App no. 19009/04 [1995] ECtHR; Kaya v Turkey, App no. 22729/93 [1998] ECtHR, para 86; Yasa v Turkey, App no. 22495/93 [1998] ECtHR, para 98; Dikme v Turkey, App no. 20869/92 [2000] ECtHR, para 101. Regarding Article 3 see A and B v Croatia, App no. 7144/15 [2019] ECtHR, paras 110 and 129 and MP and Others v Bulgaria, App no. 22457/08 [2011] ECtHR, para 111.
[47] See for example, X and Others (GC), App no. 22457/16 [2021] ECtHR; Fernandes de Oliveira v Portugal (GC), App no. 78103/14 [2019] ECtHR; Sargava v Estonia, App no. 698/19 [2021] ECtHR; Jokela v Finland, App no. 288856/95 [2002] ECtHR.
[48] Kukhalashvili and Others v Georgia, App nos. 8938/07 and 41891/07 [2020] ECtHR, para 148; Tanis and Others v Turkey, App. no 65899/01 [2008], para 160; Tagayeva and Others v Russia, App no. 26562/07 [2017] ECtHR, para 586.
[49] See, Mahmut Kaya v Turkey, App no. 22535/93 [2000] ECtHR; Kilic v Turkey, App no. 22492/93 [2000] ECtHR; and Gongadze v Ukraine, App no. 34056/02 [2005] ECtHR.
[50] Note, that in the case of Nachova and Others v Bulgaria there is a Chamber and Grand Chamber judgment. Where mentioned, we indicate to which one we are referring.
[51] The present assessment is made solely on the basis of the publicly available judgments in which the applicant’s Roma origin is expressly mentioned (even if not in connection with a discrimination claim). It is likely that there are instances where the original application or observations refer to the applicants’ Roma origin, yet this does not appear in the final judgment and therefore these cases do not appear in our selection.
[52] Interviews conducted with human rights lawyers by Emma Várnagy on 28 March 2023 (online) and 30 May 2023 for her dissertation tentatively titled ‘Unrecognized Rights – The consequences of lack or neglect of evidence of discrimination in anti-Roma police violence cases’ (interview subjects remain anonymous in the research framework).
[54] For example the case of Stoica v Romania, App no. 42722/02 [2008] ECtHR concerns a beating in front of a bar (para 8); the applicant in Vasil Shashov Petrov v Bulgaria, App no. 63106/00 [2010] ECtHR was shot in a back garden (para 6); the case of Kovács v Hungary, App nos. 21314/15, 21316/15, 21317/15 and 21321/15 [2019] ECtHR concerns an ID check at a petrol station (para 8).
[55] Borbála Kiss v Hungary, App no. 59214/11 [2012] ECtHR; Lakatosová and Lakatos v Slovakia, App no. 655/16 [2018] ECtHR; Lingurar and Others v Romania, App no. 5886/15 [2018] ECtHR; Pastrama v Ukraine, App no. 54476/14 [2021] ECtHR; Fedorchenko and Lozenko v Ukraine, App no.387/03 [2012] ECtHR.
[56] Petropoulou-Tsakiris v Greece, App no. 44803/04 [2007] ECtHR; Lingurar and Others (n 55); RR and RB v Slovakia, App no. 20649/18 [2020] ECtHR; Ciorcan v Romania, App nos. 29414/09 and 44841/09 [2015] ECtHR; Dzeladinov v FYRM, App no. 13252/02 [2008] ECtHR.
[58] Lakatosová and Lakatos (n 55); Fedorchenko (n 55); Seidova v Bulgaria, App no. 310/04 [2010] ECtHR; Lacatus (n 57).
[59] MB and Others v Slovakia no. 1, App no. 45322/17 [2021] ECtHR; PH v Slovakia, App no. 37574/19 [2022] ECtHR; Boaca and Others v Romania, App no. 40355/11 [2016] ECtHR; Mata v Hungary, App no. 7329/16 [2022] ECtHR; Kovács v Hungary (n 54).
[60] Balogh v Hungary, App no. 47940/99 [2004] ECtHR; Bekos and Koutropoulos v Greece, App no. 15250/02 [2005] ECtHR; AP v Slovakia, App no. 10465/17 [2020] ECtHR; Stefanou v Greece, App no. 2954/07 [2010]; X and Y v North Macedonia, App no. 173/17 [2020]; MB and Others v Slovakia no. 2, App no. 63962/19 [2023] ECtHR.
[61] PH v Slovakia (n 59); Eremiasova and Pechova v the Czech Republic, App no. 23944/04 [2012]; Kleyn and Alexandrovich v Russia, App no. 40657/04 [2012] ECtHR; Ognyanova and Chioban v Bulgaria, App no. 46317/99 [2006] ECtHR.
[64] Pastrama v Ukraine (n 55), paras 74–76. The Court considered this case under Article 8 instead of the original submission of Article 3.
[68] Soare v Romania, App no. 24329/02 [2011] ECtHR (racist element not investigated); in Lingurar v Romania (n 55), the applicants complained that they had been held for questioning at the police station for several hours without food or water when giving testimony about the death of their relative (the main Article 2 complaint).
[69] Anguelova v Bulgaria, App. no. 38361/97 [2002] ECtHR (6 votes to 1 no violation); Balogh (n 60), para 79; Karagiannopoulos v Greece, App. no. 27850/03 [2007] ECtHR, para 79; Vasil Sashov Petrov (n 54), para 72; Mizigárová v Slovakia, App. no. 74832/01 [2010] ECtHR, paras 122–3; Soare (n 68) (4 votes to 3 no violation’; Ion Balasiou v Romania, App. no. 70555/10 [2015] ECtHR, para 138.
[70] S Stavros, ‘Victims of racist crime: How well are their procedural rights protected under the ECHR?’ in R Spano et al. (eds.) Fair Trial: Regional and International Perspectives (Anthemis, 2020), 628.
[71] E.g. M Dembour, ‘In the Name of the Rule of Law: The ECtHR’s Silencing of Racism’ in K Bhambra & R Shilliam (eds.) Silencing Human Rights (Palgrave Macmillan, 2009), 184–202; R Rubio-Marin & M Möschel, ‘Anti-Discrimination Exceptionalism: Racist Violence before the ECtHR and the Holocaust Prism’ (2015) 26 European Journal of International Law, no. 4, 881–899.
[72] The Court has repeatedly held that ‘in so far the applicants have relied on general information about police abuse of Roma (…), the Court cannot lose sight of the fact that its sole concern is to ascertain whether in the case at hand the treatment inflicted on the applicants was motivated by racism’ (see Nachova and Others v Bulgaria, GC cited above, para 155); Bekos (n 60), para 166.
[73] J Mačkić, Proving discriminatory violence at the European Court of Human Rights (Leiden, 2017); M O’Boyle, ‘Proof: European Court of Human Rights (ECtHR)’ in The Max Planck Encyclopedia of International Procedural Law (Oxford University Press, 2017); J Gunn, ‘Limitations Clauses, Evidence, and the Burden of Proof in the European Court of Human Rights’ (2020) 15 Religion and Human Rights, no. 1–2, 199; M Möschel, ‘Is the European Court of Human Rights’ Case Law on Anti-Roma Violence “Beyond Reasonable Doubt”?’ (2012) 12 Human Rights Law Review, no. 3, 479–507.
[74] For an example of this approach in action see the leading authority in ARPV cases, Nachova (GC) (n 9), paras 162–168.
[75] Petropoulou-Tsakiris (n 56); RR and RB (n 56); Kovács (n 54); MF v Hungary, App. no. 45855/12 [2017] ECtHR; Boaca and Others v Romania, App. no. 40355/11 [2016] ECtHR; Stoica (n 54).
[76] E Várnagy, ‘“Purely Gypsy Behaviour”: Interpreting Negative Stereotypes in Racist Police Violence Cases at the European Court of Human Rights’ (2024) 6 Critical Romani Studies, no. 1, 4–23.
[77] By contrast with the cases mentioned in the case of Soare (n 68) the word ‘Gypsy’ in itself was not sufficient to trigger a separate investigation.
[78] In fact the Court has explicitly stated that while it ‘cannot exclude the possibility that in certain cases of alleged discrimination it may require the respondent Government to disprove an arguable allegation of discrimination and – if they fail to do so – find a violation of Article 14 of the Convention on that basis. However, (…) such an approach would amount to requiring the respondent Government to prove the absence of a particular subjective attitude on the part of the person concerned.’ Nachova GC (n 9), para 157.
[79] A Timmer & L Peroni, ‘Vulnerable groups: The promise of an emerging concept in European Human Rights Convention Law’ (2013) 11 ICON, no. 4, 1056–1085.
[81] ibid, and for more detail on the contrast with education cases see e.g. Timmer & Peroni (n 79) and for the contrast with domestic violence see e.g. Möschel (2012) (n 73).
[88] K Henrard, ‘The European Court of Human Rights and the “Special” Distribution of the Burden of Proof in Racial Discrimination Cases’ (2023) 4 European Convention on Human Rights Law Review, no. 4, https://doi.org/10.1163/26663236-bja10065, 426–446.
[89] M Dembour, ‘Why the European Court of Human Rights Finds It So Difficult to Acknowledge Racism’ in K Clarke & M Goodale (eds.) Mirrors of Justice: Law and Power in the Post-Cold War Era (CUP, 2009), 45–66; Rubio-Marin & Möschel (2015) (n 71); B Havelková, ‘Judicial Scepticism of Discrimination at the ECtHR’ in T Khaitan & H Collins (eds.) Foundations of Indirect Discrimination Law (Hart Publishing, 2018).
[91] R O’Connell, ‘Cinderella comes to the Ball: Art 14 and the right to non-discrimination in the ECHR’ (2009) 29 Legal Studies, no. 2, 215. See e.g. in Mata v Hungary, App. no. 7329/16 [2022] ECtHR; Balkasi and Others v Albania, App. no. 14800/18 [2022] ECtHR; Carabulea v Romania, App. no. 45661/99 [2010] ECtHR.
[92] See, among others, e.g. ERRC Third party interventions in the cases of Panayotopoulos v Greece at para 25 submitted on 5 October 2021; TK and Others v Slovakia at para 26 submitted on 20 December 2019; Balkasi and Others v Albania at para 16 submitted on 7 November 2018; AP v Slovakia at para 34 submitted on 1 December 2017; and Pastrama v Ukraine at para 24 submitted on 10 November 2016.
[98] In Lakatosová and Lakatos (n 55), the Court examined the claims as Articles 14 and 2 only and found a procedural violation thereof. The cases in which there is no separate holding on the core Article are indicated as neither substantive nor procedural in Figure 1.
[99] Kleyn and Aleksandrovich (n 61); Fedorchenko (n 55); Mihaylova and Malinova v Bulgaria, App. no. 36613/08 [2015] ECtHR.
[107] Jasar v FYRM, App. no. 69908/01 [2002] ECtHR; Dzeladinov and Others v FYRM, App. no. 13252/02 [2008] ECtHR; Sulejmanov v FYRM, App. no. 69875/01 [2008] ECtHR; Durdevic v Croatia, App. no. 52442/09 [2011] ECtHR; Ciorcan (n 56); Ion Balasiou (n 69); Adam v Slovakia, App. no. 68066/12 [2016] ECtHR; Gheorghita and Alexe v Romania, App. no. 32163/13 [2016] ECtHR; MB no. 2 (n 60); Balkasi and Others v Albania, App. no. 14800/18 [2022] ECtHR.
[111] Ion Balasiou (n 69), para 120; Ciorcan (n 56), para 115; MB and Others v Slovakia no.1 (n 59), para 60.
[113] ibid, paras 116–123. For further discussion of these types of cases see, for example, Gerards (n 21).
[115] Velikova v Bulgaria, App. no. 41488/98 [2000] ECtHR; Anguelova (n 69); Balogh (n 60); Ognyanova and Choban v Bulgaria, App. no. 46317/99 [2006] ECtHR.
[120] In both of these cases the Court relied on explicit racial profiling and slurs uttered by the abusive officers which met the beyond reasonable doubt standard of proof. See Stoica (n 54), para 76 and Lingurar (n 55), para 122. While these judgments are laudable, they stand to highlight that only very explicit circumstances warrant a finding of a substantive violation of Article 14 whereas in reality the very nature of biases is a lot more elusive.
[121] Nachova [GC], (n 9); Bekos and Koutropoulos (n 60); Cobzaru v Romania, App. no. 48254/99 [2007] ECtHR; Petropoulou-Tsakiris (n 56); Ciorcan and Others v Romania, App. nos. 29414/09 and 44841/09 [2015] ECtHR; Boaca and Others (n 59); Lingurar (n 55); MF (n 75); Kovács (n 54); MB and Others no. 2 (n 60).
[122] Nachova [GC], (n 9) para 163; see further e.g. Bekos (n 60), para 72; MB no. 2 (n 60), para 95.
[124] RR and RD v Slovakia, App. no. 20649/18 [2020] ECtHR; MB no. 2 (n 60); MF (n 75); Lingurar (n 55); Boaca (n 59); Cobzaru (n 121); Bekos (n 60); Balogh (n 60).
[125] See, among others, ERRC Third party interventions in the cases of Panayotopoulos v Greece at para 25 submitted on 5 October 2021; TK and Others v Slovakia at para 26 submitted on 20 December 2019; Balkasi and Others v Albania (n 107), para 16 submitted on 7 November 2018; AP v Slovakia at para 34 submitted on 1 December 2017; and Pastrama v Ukraine (n 55), para 24 submitted on 10 November 2016.
[126] X and Y v North Macedonia (n 60); MB no. 1 (n 59) due to lack of information indicating biased motives); AP v Slovakia (n 60) due to unsubstantiation; Fogarasi and Others v Romania, App. no. 67590/10 [2017] ECtHR; Adam (n 107); no prima facie case of discrimination established; Borbala Kiss, App. no. 59214/11 [2012] ECtHR, no evidence disclosing discriminatory conduct.
[127] See e.g. E Várnagy, ‘Bódi And Others V Hungary: When The Court’s Focus On The Volume Of Procedures Speaks Volumes About Its Stance On Antigypsyism’ (2023) Strasbourg Observers, https://strasbourgobservers.com/2023/09/15/bodi-and-others-v-hungary-when-the-courts-focus-on-the-volume-of-procedures-speaks-volumes-about-its-stance-on-antigypsyism/; and in case law: Balázs v Hungary, App no. 15529/12 [2015] ECtHR; Sabalić v Croatia, App no. 50231/13 [2021] ECtHR, paras 106–9; Škorjanec v Croatia, App. no. 25536/14 [2017] ECtHR, paras 58 and 68; Lakatosová and Lakatos (n 55), para 74.
[128] Regrettably, the Court states that ‘the Government cannot be required to prove the absence of a particular subjective attitude on the part of the person/persons concerned’ (see Nachova and Others GC, cited above at n 9, para 157) which it reiterates in its non-violent discrimination case law too, such as in the recent racial profiling case of Muhammad v Spain, App. no. 34085/17 [2022] ECtHR, para 95.
[130] Timurtaş v Turkey, App. no. 23531/94 [2000] ECtHR; Taş v Turkey, App. no. 24396/94 [2000] ECtHR; Conka v Belgium, App. no. 51564/99 [2002] ECtHR. This is discussed in some detail in K Kamber, ‘Substantive and Procedural Criminal Law Protection of Human Rights in the Law of the European Convention on Human Rights’ (2020) 20 Human Rights Law Review, no. 1, 75–100.
[134] ERRC Third party interventions in the case of TK and Others v Slovakia at para 26 submitted on 20 December 2019; AP v Slovakia at para 34 submitted on 1 December 2017.
[137] C Roberts, ‘Reversing the burden of proof before human rights bodies’ (2021) 25 The International Journal of Human Rights, no. 10, https://doi.org/10.1080/13642987.2020.1859486, 1687.
Acknowledgements
We are grateful for the organizers of the 2023 Institutions for Conflict Resolution (COI) Conference ‘Towards just institutional approaches to conflict prevention and resolution’ for creating a supportive and inspiring environment, and all the participants of our session for their valuable feedback on our paper. We are also thankful for the feedback we received from the participants of the 2023 International Interdisciplinary and Transdisciplinary Conference ‘The Universal Declaration of Human Rights at 75: Rethinking and Constructing its Future Together’ organized by the Human Rights Research Network at Ghent University. We thank the reviewers for their insightful suggestions. Any errors remain ours.
Funding Information
This paper draws upon the authors’ research conducted within the framework of the IBOF research project ‘Future-proofing Human Rights: Developing Thicker Forms of Accountability’ at Ghent University, Hasselt University, the University of Antwerp, and the Vrije Universiteit Brussel (grant ID: Special Research Fund iBOF/21/031).
Competing Interests
The authors have no competing interests to declare.
Author Contributions
The paper constitutes the outcome of the authors’ common reflections. However, Sections 2 and 4 have been drafted by Harriet Ní Chinnéide, while Section 3 and the table illustrating the case law can be attributed to Emma Várnagy. Sections 1, 5 and 6 have been jointly drafted by both authors.