2.1 Introduction
States violate human rights. If this were not the case, a multi-layered human rights framework aiming at accounting for states’ human rights violations would not have been created. States may be participating in this accountability system in good faith, by which we mean they are willing – at least to some extent – to acknowledge their behaviour and change their ways. In other situations, however, human rights violations are deliberate and entrenched, having become part and parcel of state policy. Examples include systematic practices of torture, forcible disappearance, assassination of political opponents or, coming to the subject of our chapter, recklessly and often violently pushing people on the move back over the state border. A state which intentionally commits serious violations of human rights is unlikely to own up to its actions. Its tendency will be to deny the violations – ensuring at the same time that as little evidence of them as possible emerges. The resulting evidentiary gaps then make it difficult for accountability mechanisms to operate as intended.
Successful litigation should be possible even in these circumstances. However, it depends on the adjudicator accepting a complaint’s factual basis. If states suppress evidence of their deeds, conclusive evidence that a complainant was treated as alleged will be lacking. Still a bedrock in evidentiary law today, the Latin adage actori incumbit probatio (‘it is for the claimant to prove their claim’) is often taken to mean that a complaint left without the proof of its factual allegations stands to be rejected. Such an outcome is problematic and requires a corrective when the complainant’s evidentiary incapacity stems from actions or omissions purposefully instigated by the perpetrating state precisely in order to avoid human rights accountability.
One solution is to have the adjudicator shift the burden of proof from the complainant to the respondent state.Footnote 1 Although such a shift is often referred to in legal scholarship, the question of its exact conditions of operation is not settled. Surprisingly human rights jurisprudence tends not to address it head-on. In-depth academic analyses are conspicuous mostly by their absence.Footnote 2 We shall propose that the shift be conditioned on two sets of circumstances being established, each attracting a different standard of proof. We shall further argue that the shift-based evidentiary route must be distinguished from the ‘normal’ evidentiary route whereby, against the odds of state-created holes, the facts can get established on the substantive strength of the evidence submitted.
To make these arguments, we use the adjudication of pushback complaints by UN treaty bodies (UNTBs) as a case study. Pushbacks are expulsions designed to circumvent the law, by the state, of people on the move. (We favour this expression over other terms such as ‘migrants’, due to its bypassing the problematic migrant/refugee binary at the heart of most legal frameworks and migration discourses).Footnote 3 When the state ‘pushes’ people ‘back’, it forces them out of its jurisdiction, in such a way that they cannot access procedural and/or other legal safeguards.Footnote 4 Such evasion of legal frameworks may concern international instruments, such as the Refugee Convention or the protection granted to a person who is, for example, a victim of torture, a trafficked victim or an unaccompanied minor. It may also concern provisions found in domestic and international law for the benefit of the whole population, such as those adopted in order to guard against arbitrary detention or excessive state violence.
The evasion of legal guarantees is what makes an expulsion a ‘pushback’ (without the latter word being a legal term). This means that a pushback always entails expulsion-related human rights violations, including the denial of the expelled’s right-holding status.Footnote 5 In addition, between the moment of their apprehension by state agents and their expulsion, people may have been beaten, secretly detained, tortured, left to die, disappeared or refouled to a place where their life or integrity is at risk – in short, subjected to serious human rights violations.Footnote 6
In Europe, one would hope the European Court of Human Rights (ECtHR or Court) to be the go-to body for individuals seeking redress for the violations of human rights they have suffered. However, as Dembour has extensively documented,Footnote 7 the Court has shown itself reluctant to defend the human rights of people on the move. This observation also holds for pushbacks: the Court has hardly ever found a ‘no-paper pushback’Footnote 8 – that is one that leaves no trace of their occurrence in state records, as explained below – to have breached the European Convention on Human Rights.Footnote 9 The orientation of its jurisprudence is inciting litigators, including the European Centre for Constitutional and Human Rights (ECCHR), for which Hakiki works, to address pushback complaints to UNTBs, a telling change of forum since these bodies are not empowered to deliver legally binding judgments.
UNTBs’ final views are nonetheless precious: carrying some authority, they offer some redress to victims as well as contribute to the development of human rights international law. In respect to our topic, they provide an official confirmation, against state denials, that pushbacks are a reality. Our analysis of the four final views UNTBs have so far adopted in pushback casesFootnote 10 concludes that the correct factual determinations were reached, with the UNTBs willing to see through the state-instigated evidentiary holes, but on a weak legal reasoning. In response, we identify evidentiary principles which we argue can contribute to the building of a fair, clear and coherent evidentiary framework.
Our chapter proceeds as follows. Section 2.2 classifies pushbacks into ‘no-paper’ and ‘paper’ categories, with each type giving rise to particular forms of state denials. Section 2.3 reviews the limited evidentiary elements which, in a context of state-created evidentiary holes, may nonetheless potentially remain accessible to a pushback survivor. The chapter then moves to a legal analysis. Section 2.4 starts by mapping out the four logical outcomes which the adjudicator can reach regarding the disputed facts of a complaint: these can be upheld or rejected, procedurally or substantively so, following a number of evidentiary principles. This conceptual exercise provides the basis for critically analysing, in Section 2.5, the four pushback cases which have been decided by UNTBs to this day. Section 2.6 identifies what we regard as the two major risks of a loose evidentiary reasoning: the adjudicator’s failure to demarcate the procedural (shift-based) and substantive (evidence-based) evidentiary routes; the upscaling of applicable standards of proof (e.g., due to the higher-than-legally-required standard met in some extraordinary cases becoming expected in subsequent cases, even where the shift is called for). Section 2.7 concludes.
2.2 State Denials: Typical Forms in ‘No-Paper’ and ‘Paper’ Pushback Cases
The stronger the secrecy surrounding a pushback operation, the less effective the legal safeguards against abuse, the more serious the human rights violations tend to become. This spiral then provokes states to want to deny their actions and in turn to dispense huge efforts in suppressing evidence. The following measures have been reportedFootnote 11 to have been taken by states with regard to pushbacks: systematically failing to record people’s presence, handling and custody; using unmarked vehicles for transportation, unofficial means for communication and unofficial detention sites; redirecting border cameras to ‘unwitness’Footnote 12 human rights violations or refusing to share border cameras’ footage;Footnote 13 criminalising audiovisual recording by third parties; seizing and destroying people’s mobile phones; and expelling victims and witnesses, thus making them inaccessible to investigations.Footnote 14
In the course of litigating cases, the ECCHR has come to distinguish between two broad types of pushbacks. No-paper pushbacks leave no trail in official records, such that even the most minutious examination of state records would give no hint whatsoever that they ever took place. By contrast, paper pushbacks leave an official trace, for example a refusal-of-entry decision or a readmission under a bilateral agreement, both of which qualify as pushbacks if the authorities ignored the protection needs and rights of the person refused or ‘readmitted’.
The ECCHR has observed that when a person complains to have been no-paper pushbacked, their apprehension by state agents – of which there is by definition no official record – is often the centrally disputed fact. The situation presents itself differently in paper pushbacks, where the core problem is that the state record typically fails to mention protection claims. The factual issue most saliently disputed in this type of complaint is not whether the person was expelled (this is clearly documented) but whether they had expressed, or could have expressed, protection needs at the time of their expulsion. In paper pushbacks, the state typically claims to have respected all relevant applicable legal frameworks.
2.3 Which Types of Evidence Are Accessible to a Pushback Survivor?
Whilst the complainant is generally expected to bring the proof of their factual allegations, state-created evidentiary holes make this difficult, if not impossible. Based on Hakiki’s litigating experience, this section identifies nine types of evidentiary material a pushback victim may nonetheless potentially be able to submit, as follows: (1) the complainant’s own account; (2) identity documents; (3) digital messages and pictures/videos; (4) medical records; (5) fellow survivors’ testimonies; (6) technological analysis of digital material; (7) legal correspondence; (8) state documents; (9) general reports. This section briefly discusses each element, highlighting their challenges and potential value. Elements (7) and (8) are especially relevant to paper pushbacks, whilst the others are addressed having mostly no-paper pushbacks in mind.
2.3.1 The Complainant’s Own Account of the Events
With rare exceptions, pushback complainants have little to no external, stand-alone conclusive evidence of their factual claims. In this context, their own account of what happened to them often becomes the central piece of evidence they submit to the adjudicator. Few survivors, however, are ready to go through the process of recounting their pushback, as this tends to be a lengthy and retraumatising experience, often affecting a person for days and even weeks.
If a survivor wants to enter this process – which they should be encouraged to do only if they are already in a stable environment, ideally with access to psycho-social services – they cannot be expected to write their account alone. That they will need assistance is recognised in adversarial legal systems, which, leaving it to the parties to produce evidence,Footnote 15 devote a substantial part of the training of lawyers-to-be to the taking of truthful, non-leading and clarifying statements. When the witness is a survivor, additional specialised training is required so as to minimise the re-traumatisation which occurs when a person is brought back to their original ordeal.Footnote 16
A ‘detailed and credible’ account (to cite qualificatives found in the jurisprudence) should attract high probative value, given it is virtually impossible for a person who has not directly lived a situation to recount it without getting into ‘twists’ and introducing incoherences within their previous inventions. When such an account is submitted, the evidentiary elements discussed below mainly serve to corroborate it. By contrast, these elements have to become the main evidence when a victim cannot recall their pushback in detail, possibly due to memory loss provoked by the trauma of the pushback.
2.3.2 Identification Documentation
States commonly object in pushback litigation that the complainant is not who they say they are. Even though they most likely will have travelled undocumented and with their presence in most of the countries they crossed left unregistered, the complainant will typically have in their possession some documents bearing their identity, which they can submit to the adjudicator. These papers, however, will often have been hastily drawn by the authorities, without respecting specific spelling or transcription in the Latin alphabet. The state then commonly relies on orthographic variances to argue that the complainant is not the same person as, for instance, the person whose readmission from one state to another was registered. Proving the basic fact of one’s identity becomes a challenging evidentiary matter for the complainant, necessitating corroboration and/or proof of credibility.
2.3.3 Digital Messages and Pictures/Videos Sent Around the Time of the Pushback
People often lose their belongings or have them destroyed, including by the police, during their journeys. This is particularly true of mobile phones, as already mentioned. Thus, original footage with metadata – including geolocation of where it was recorded – will rarely be available.Footnote 17 However, important evidence may be retrieved from messaging services and social media platforms, which remain accessible even after losing access to the mobile phone that was originally used. Proofs are also created from keeping supportive contacts (kin, activists, lawyers, journalists, non-governmental organisations [NGOs]) informed of one’s whereabouts and state of health through sending messages, including videos and/or pictures. These may serve during litigation to prove that the complainant was at the place they say in their submissions and are not inventing the treatment their complaint alleges they have suffered. In awareness that states are prone to subsequently deny the abusive treatment they inflict, a victim occasionally seeks to record what is being done to them as it is being done, in which case they are likely to then ask trusted contacts to safeguard electronically sent information (see examples in Section 2.3.6).
A technical note is in order. Posts and messages on social media platforms such as Facebook, Instagram or TikTok theoretically remain accessible even after phones are lost or destroyed. By contrast, messages sent through messaging services such as WhatsApp, Telegram or Signal tend to get lost on the sender’s end, but may be accessible if receivers safekept them (for instance in the home country). The metadata of a file shared through either technology is automatically erased and replaced by a new time-record indicating the day the file was last shared. However, the evidentiary significance of the absence of original metadata should not be overestimated, in light of the limited probative value of metadata, especially given the possibility to attribute later dates to pictures (for instance by taking a picture of a picture). In extremely rare circumstances, applicants may be able to safekeep original pictures or videos, with the original metadata, sometimes even including the GPS location of the taking of the picture or video. However, because of the routine seizure and destruction of phones, this is rare.
2.3.4 (Absent) Medical Records
In instances of violence and injury, medical records will rarely exist, since people on the move have little to no access to medical assistance. Some victims may photograph their injuries, though the resulting pictures will rarely be dated, as addressed above, let alone comply with forensic requirements. For example, they may show the injured body parts but not the complainant’s face, or may lack scale references to understand the extent of the injury, potentially requiring additional expert analysis.
2.3.5 Fellow Survivors’ Testimonies
Whilst on the move, people form temporary communities whose members are, like them, undocumented and vulnerable. As people tend to travel and be pushed back in groups, a complainant should theoretically be able to rely on the testimony of fellow survivors. This could be extremely precious in the (often state-instigated) absence of local civilian witnesses. However, most complainants will have lost contact with their travel companions by the time they submit a claim. Further, even if an erstwhile companion can be located, they may not wish to provide a statement.
Where a willing survivor is identified, the collection of their testimony is logistically demanding. People who are pushed back together rarely form a homogenous group: they generally come from a mixture of countries, speak different languages and end up in various places. Obtaining a testimony therefore typically necessitates transnational travel and translation into several languages. The task is also emotionally demanding: it is as painful, intense and laborious a process, and requires as much training, experience, empathy and resources, as taking a statement from a complainant. This is because ‘witnesses’ (in the proceedings) are themselves victims/survivors, often of several pushbacks. In particular, it would be wrong to ask them to focus strictly on the elements relevant to the complaint being litigated, as preventing them from recounting what is important to them could aggravate their trauma.Footnote 18
Procedurally, a statement from a witness presents the advantage that adjudicative bodies are familiar with this type of evidence.Footnote 19 What is more, the circumstances of its production may make it immediately clear to the adjudicator that the complainant and the witness have had no recent interaction, thereby eliminating the possibility of them having ‘concocted’ the statement, thereby enhancing the probative value of the submitted testimony.
2.3.6 Technological Analysis of Digital Materials by Organisations Such as Forensic Architecture
When an exceptional level of evidence is gathered in support of the victim’s account, this may provide the opportunity for its evidential value to be further enhanced through technological operations. This section recounts two examples drawn from the ECCHR case work where this happened.
Our first example is the case of Parvin A., an Iranian refugee who was pushed back from Greece to Turkey no less than six times.Footnote 20 Despite the risks, Parvin A. had resolved to hide and use her phone for recording the circumstances of her apprehensions and detentions during the first pushbacks she suffered. Through messaging services, she sent out over forty pieces of audio and/or visual material to trusted contacts. Forensic Architecture (FA), the University of London-based research group famous for its pioneering use of cutting-edge techniques for the investigation of state and corporate violence, worked on this material. It reconstructed the events complained of in a platform specifically designed to this effect, even creating a 3D model of the detention sites where Parvin A. had been placed during her first pushback.Footnote 21 Images FA subsequently obtained through open sources led to the identification of the police stations in which she had been secretly detained. They corroborated her account regarding inter alia the way people were handled. The similarities between her and other pushback cases involving the same police stations and similar forces revealed the scope of coordination and infrastructure deployed by the state for pushback operations, confirming her treatment was not an ‘isolated event’ but part of a ‘pattern’ or ‘practice’ by the Greek authorities.Footnote 22
Our second example is the case of I. M., a refugee from the Pakistani-controlled side of Kashmir, who was pushed back from Slovenia to Croatia, and severely ill-treated in Croatia before being pushed back to Bosnia and Herzegovina (BiH). A courageous travel companion had taken a video recording after I.M.’s ill-treatment and expulsion to BiH, showing him in great pain. The video was analysed and verified by the NGOs No Name Kitchen/Border Violence Monitoring Network, and the case was investigated again by journalists from Lighthouse Reports and German media outlet Der Spiegel.Footnote 23 Its originally fuzzy imagery was stabilised, leading to the identification of (1) the location of the incident through GPS metadata and corroboration from a close examination of the landscape; (2) the affiliation of the officers appearing in the footage through an analysis of the details of their uniforms; and (3) the identification of the perpetration of similar acts of violence and expulsions in the background of the footage.
The afore-described technological feats are evidentiary gold. However, they are extremely costly and unavailable to most pushback victims, the overwhelming majority of whom do not even have access to legal representation. Three or more organisations belonging to the not-for-profit and/or journalistic worlds worked on each case for months. The dispensation of such a level of resources clearly cannot set the bar for what an adjudicator may expect to see evidenced in any given case.
2.3.7 Legal Correspondence
Rarely, people on the move may be in contact with lawyers as they are literally facing pushback risks at the border. These lawyers may represent them before the authorities, for example faxing international protection claims to national border authorities to put the latter on notice of their clients’ asylum-seeking status, or seeking interim measures from regional and/or international adjudicative bodies such as UNTBs, despite risks of criminalisation, for example, for facilitating irregular movement. Correspondence which has arisen between lawyer, state and international body may later be submitted to defeat the state’s claim that the complainant had never expressed or intended to make a protection claim at the border.Footnote 24
2.3.8 State Documents on the Complainant’s Case
In paper pushbacks, the state will per definition have documented the supposedly legitimate measure through which it has pushbacked the complainant. These state documents may constitute the sole evidence the complainant is able to submit to prove their interaction with state authorities. Needless to say, the state will have had exclusive control of what they record. Their content, therefore, including the absence of any protection claim, should be approached with caution by the adjudicator when assessing the content of said interaction.
2.3.9 External Reports Documenting State Practice
Published reports documenting a state practice of pushbacks are an essential part of the complainant’s submissions. They make sense of the evidentiary difficulties which mar the complaint. Moreover, they usually confirm some of the details provided in the complainant and/or the witnesses’ accounts, strengthening their credibility. The gathering and submission of reports from as authoritative sources as possible is also a highly time-consuming task.Footnote 25
2.3.10 Conclusion
This section has identified the types of evidentiary materials that a pushback complainant may be able to submit to an adjudicatory body in support of their factual allegations. Our review has made clear that for a complainant to submit strong evidence necessitates the deployment of resources to which most people on the move – and their lawyers – do not have access. Expecting pushback complainants to routinely submit strong evidence would defeat human rights accountability and justice. In a context where the state is likely to dispute whatever weak or strong evidentiary submissions the complainant is able to make, an in-depth reflection as to how the evidentiary system should work is needed. Section 2.4 starts this work by abstractly mapping out the possible end forms of a factual dispute.
2.4 Disputed Facts: Four Logically Possible Adjudicatory Outcomes
Pushback litigation sees the word of the complainant pitted against state denials: the facts are thus ‘in dispute’ between the parties. Such a situation is not one that international human rights law relishes; it prefers when the facts of a case arrive before the international adjudicator already established by the domestic authorities. Pushback victims, however, by definition have no opportunity to challenge their expulsion prior to being pushed back, and they also usually face a system of utter impunity with regard to any ill-treatment or arbitrary detention, with no possibility of having their claims effectively investigated and dealt with within the national criminal system.Footnote 26 In short, no domestic court will have properly considered their case, if at all. The admissibility requirement of exhaustion of domestic remedies, amply discussed elsewhere in this volume, should consequently be waived,Footnote 27 leaving it to the international judge to determine the facts of the complaint.
Abstractly thinking about how a disputed fact can present at the end of the adjudicatory process, two outcomes are logically possible: either it remains disputed and unclear, or it has been resolved – each time in favour of either the complainant or the respondent state.Footnote 28 This results in four logical possible factual outcomes, as per Table 2.1:
Table 2.1 Long description
The data are as follows.
Facts remain unclear, to the benefit of the state, Outcome 1.
Facts remain unclear, to the benefit of the complainant, Outcome 2.
Facts are resolved, in favor of the complainant, Outcome 3.
Facts are resolved, in favor of the state, Outcome 4.
In practice, things are unlikely to be as clear-cut as Table 2.1 makes out. For example, one disputed point may be resolved in favour of the complainant, another in favour of the respondent, whilst others remain unresolved. Importantly, it is not because one party’s factual allegations are accepted as established that this party will win the case – they may lose on legal arguments. Finally, whether a particular judicial decision would best be categorised as supporting one or the other party’s account is not always obvious. Table 2.1 must be approached as a heuristic device.
Outcome 1 arises when a complainant fails to establish the factual basis of their complaint. It can thus be conceptualised as an application of the actori incumbit probatio principle. It is important to underscore, however, that the Latin term actor (used in its dative form actori’) is really meant to designate any party who is making a factual claim in the proceedings so that the adage does not so much allocate the burden of proof to the complainant as distribute it between the parties (even if in practice the initial burden rests on the complainant). Outcome 1, therefore, need not be the end point for all the cases whose facts remain uncertain by the end of the adjudicatory process – even under the adage.
Outcome 2 arises when, despite the continual lack of factual clarity, the adjudicator eventually considers the complainant’s factual allegations established, due to having shifted the burden of proof from the complainant to the state, and the latter not having then discharged it. Although the shift (also known as ‘reversal’) in the burden of proof is much talked about in international human rights adjudication, its exact conditions of operation have not been clarified. We submit that the shift should be subject to two conditions, which we propose to call the context condition and the linkage condition. We further submit that each of these conditions should be proven to a different standard of proof, as we explain after having introduced this concept.
One need not be a lawyer to understand that a fact can be established to a higher or lesser degree of certainty – in the same way that one can either very strongly or more circumspectly believe that something has happened. The Common Law recognises four such degrees: a finding can be so secure as being ‘beyond reasonable doubt’; it can be made on evidence which is ‘clear and compelling’ (less than virtually certain, but more than probable); it may be adopted ‘on the balance of probabilities’ (being ‘more likely than not’, that is, having at least fifty-one per cent chances of being correct); and, at the bottom of this descending scale, it may be no more than just plausible – prima facie. Evidentiary rules refer to these various bands of persuasion to indicate the level – in legal language, ‘threshold’ or ‘standard of proof’ – to which a disputed factual allegation must be proven for the adjudicator to accept to be persuaded to uphold it. These are (in phrasings which in practice may be expressed under slightly different formulations):
1) Beyond reasonable doubt (BRD)
2) Clear and compelling
3) More probable than not
4) Prima facie (meaning ‘at first sight’ in Latin, and evoking a standard which is also sometimes expressed as requiring no more than a ‘beginning of proof’)Footnote 29
This Common Law classification has no equivalent in the Civil Law tradition, which contents itself with speaking of the intime conviction du juge (‘inner conviction of the judge’), without attempting to differentiate the degree or depth of the judge’s conviction. This stark difference in evidentiary approaches may well explain the rarity of references to specific standards of proof in international human rights adjudication: many international judges have not been trained in them. This, however, is a great pity. Alongside the other contributors to this volume, we have come to the conclusion that the notion of the standard of proof is an indispensable evidentiary tool to achieve fairness and justice in human rights adjudication – as the operation of the shift, to which we now return, may serve to illustrate.
The shift’s first condition commands that there exists a context which indicates that the respondent state may well have acted in the way claimed by the complainant. Roberts distinguishes two such broad contextual situations: (1) the state is known to have previously been acting in the way described by the complainant (state practice or pattern), or (2) the state is refraining from sharing evidence which it either possesses (due, e.g., to the events having occurred under state surveillance) or should possess (due, e.g., to a legal duty on the state either to record the contested event, as for any detention, or to investigate, as in the case of serious injury, death or disappearance, the more so when the latter occurs under its custody).
We concur with Roberts that the context condition should neither have to be established beyond reasonable doubt nor be found established merely on prima facie evidence. Roberts proposes that the general standard applicable in human rights adjudication, namely more probable than not, should be applied here too. We believe some adjudicators may be more comfortable to impose the clear and compelling standard in this instance. Of course, in practice, whether the clear and compelling or the probable standard is reached is a judgment call. Whatever standard is adopted, the important thing is that, in case of hesitations, the adjudicator favours finding the condition established rather than rigidly insisting that the complainant submits a level of evidence which is beyond their reach.Footnote 30 One would expect the context condition to be established in respect to virtually every pushback complaint. If it is not, it may indicate that the adjudicator expected an unreasonably high level of evidence or, alternatively, that the complainant did not have sufficient resources to gather the documentation that could have established that it was not unreasonable to presume (no more than that) that the state acted in the way alleged.
The shifting of the burden of proof depends on the linkage condition also being met. This condition requires the applicant to establish that it is plausible that they were personally affected by the presumed events, in other words, that they were one of the people (in our selected topic) pushed back. Importantly, this only needs to be established prima facie, a ‘beginning of proof’ thus being sufficient for the condition to be found established. Accordingly, just one of the following pieces of evidence next to the applicant’s statement should be enough for the adjudicator to consider the linkage condition fulfilled:
1) A detailed account from the applicant of what happened to them
2) A picture of the pushback operation taken despite the obstacles, where the applicant appears
3) A witness statement emanating from a fellow migrant who testifies having seen the applicant being pushed back
4) A social media message recounting the event and the presence of the applicant in it, sent shortly after the pushback in question occurred
5) Confirmation by whatever means of the applicant’s location during, immediately before or immediately after the pushbackFootnote 31
If these two conditions are met, the adjudicator must shift the burden of proof onto the state. It therefore now falls on the state to rebut the presumption of veracity of the complainant’s factual allegations. In the absence of said rebuttal, the adjudicator must uphold the complainant’s factual account, even if the facts remain uncertain. Were this outcome not reached, it would defeat the purpose of having introduced the mechanism of the shift as a corrective to the application of the normal evidentiary rules.Footnote 32 To be noted that the standard of proof which the state’s rebuttal must meet for it to be accepted by the adjudicator is hardly ever discussed in the literatureFootnote 33 or spelt out in the case law. Our opinion is that it should be at least clear and compelling and, in some circumstances, even BRD.
In the last analysis, Outcomes 1 and 2 are procedural in nature: under Outcome 1, the complaint is ‘automatically’ rejected when the adjudicator assesses its factual basis has not been evidenced; under Outcome 2, the factual basis of the complaint must be upheld once it is seen that the state has not rebutted the presumption which has arisen that it acted in the way alleged by the complainant.
By contrast, Outcomes 3 and 4 are substantive. Like Outcome 2, Outcome 3 ends up upholding the complainant’s factual allegation, but here it is because the adjudicator finds for the complainant’s version of the facts upon the substantive consideration of all the circumstances of the case – including any presumption which arises against the state’s factual allegations/denials. The taking into account of presumptions is entirely appropriate since, under the Latin adage quoted above, the state is one of the parties to the proceedings expected to bring the proof of all their factual claims. This important consideration is too often neglected.
There is no requirement before the UNTBs for the complainant to prove their case beyond reasonable doubt.Footnote 34 Although hardly ever explicitly stated, the default standard of proof applied in international human rights adjudication at the merits stage seems to be the balance of probabilities (i.e., something being ‘more probable than not’), also sometimes expressed as ‘preponderance of evidence’ (more evidence leans on one side as opposed to the other). Consequently, whenever the adjudicator is able to find that the complainant’s version of the facts is, on balance, more persuasive than the state’s, the corrective of the shift is not needed and should not be applied. Even if the conditions for such a shift, as identified above, appear to be met, in reality they are not since the second condition is proven to a higher degree than required. In such a scenario, the adjudicator should find the facts established on their own probative basis, resulting in a substantive finding – an outcome which is preferable to a procedural finding obtained through the shift. We come back to this point later.
Finally, it is of course theoretically possible that the adjudicator comes to the conclusion that the facts are as described by the state rather than by the complainant. Outcome 4 captures this scenario in our table. As with all other outcomes, the adjudicator should explain why they reach this conclusion.
A new table (Table 2.2) can be produced which incorporates the normative discussion above.
Table 2.2 Long description
If the facts remain unclear
Outcome 1. Procedural rejection.
Legal basis. The complainant has not established the factual basis of their complaint. The latter is therefore rejected as per the Latin adage.
Outcome 2. Procedural upholding.
Legal basis: The complainant having established a suspicious state context and brought a beginning of proof that their case relates to this context, the burden of proof has been shifted by the adjudicator onto the state, which has failed to discharge it. The factual basis of the complaint is considered established, generally leading to violation findings.
If the facts are resolved.
Outcome 3. Substantive upholding.
Legal basis. The adjudicator finds the complainant's version of the facts established, on the balance of probabilities or to a higher standard. Although this factual outcome does not guarantee the complaint will be upheld on legal grounds, this normally follows.
Outcome 4. Substantive rejection.
Legal basis. The adjudicator finds the state's version of the facts established. This factual outcome does not automatically lead to the complaint being rejected on legal grounds.
Two remarks are in order. The first is an important legal point: the discussion offered above exclusively concerns scenarios which involve disputed facts. When facts are uncontested, the principle is that the adjudicator is to consider them established. The second is a reminder that neither Table 2.1 nor Table 2.2 intends to capture the complexities of the potentially multiple evidentiary processes and decisions which arise in any one case. Whilst the Tables are drawn as if one single and clear factual conclusion was reached in any given case, which would nicely encompass any and all more specific factual ‘sub-claims’, things do not necessarily work like this.
2.5 How UNTBs Have Navigated the Challenge of Pushbacks’ Disputed Facts: Case-by-Case Analysis
To our knowledge,Footnote 35 so far UNTBs have only decided four pushback complaints.Footnote 36 This small number gives us the opportunity to review each final view in detail. We do this in the chronological order of their adoption. For each case, we start by offering a summary of the complainant’s factual account and the response it attracted from the respondent state (often, denials following the patterns identified in Section 2.2), before analysing the reasoning provided by the committee on the facts. We conclude that the UNTBs have reached the correct determinations, but following an evidentiary reasoning lacking in clarity and sharpness.
2.5.1 Sonko v. Spain
Sonko, the first case to be reviewed, was decided by the Committee against Torture (CAT) in 2011.Footnote 37 Both parties agreed on the following facts: Mr Sonko and three other persons had been in Moroccan waters, each floating on a dinghy and trying to reach the Spanish enclave of Ceuta. The group of four was pulled up into a boat by the Spanish Guardia Civil (a paramilitary force) and taken to Moroccan waters, and then thrown back into the sea at a place where they could not stand up. As Mr Sonko started drowning, a Spanish officer swam to him, dragged his body onto the shore and tried to reanimate him, but to no avail – Mr Sonko had died. The parties disagreed on one crucial fact: whether the officers had pierced Mr Sonko’s dinghy before throwing him back into water.
Atypically for a no-paper pushback, Spain opened a domestic judicial investigation of the incident. Whether the complainant (the deceased’s sister) was informed of these proceedings was disputed between the parties. Spain would only have needed to produce a copy of its alleged notification for this issue to be settled. Without remarking upon this, let alone requesting the State to produce said copy, CAT stated that the domestic remedies had been exhausted (a bewildering determination since it was not the complainant who had pursued the domestic investigation, which the Committee had moreover specifically recognised to have been flawed).Footnote 38
In yet more atypical submissions in a ‘no-paper’ case, Spain neither denied before CAT that the pushback had taken place (on the contrary, it confirmed its occurrence by saying it had been carried out ‘in accordance with the procedures for dealing with immigration by sea’),Footnote 39 nor that Mr Sonko had encountered its officers. We hypothesise that the State refrained from totally denying the pushback due to the victim’s very public death in the presence of its agents and on foreign territory, which would likely have produced local knowledge and thus potential witnesses.
CAT upheld the complaint, having observed that
it [fell] to the State party to explain the circumstances surrounding Mr. Sonko’s death, considering that he was alive when he was pulled out of the water. [R]egardless of whether or not the Civil Guard officers punctured Mr. Sonko’s dinghy or at what distance from the shore he was expelled from the boat, he was placed [by the authorities] in a situation that caused his death.Footnote 40
This reasoning resonates with the principle according to which a presumption of state responsibility arises (which the state must rebut) when a person comes under its custody or control in good health, and then leaves it injured or dead. Whilst its gist is to be approved, the reasoning falls short in two respects.
Firstly, the Committee does not explicitly state that the State had failed to provide a convincing explanation for the death of Mr Sonko. This conclusion, albeit congruent with the finding that the State had breached its obligation to investigate, is left implicit.Footnote 41 Secondly, the evidentiary basis upon which the violation is found is not exactly identified. In evidentiary terms, this could have been either because the State had failed to discharge the burden of proof which the Committee had shifted onto its shoulders (Outcome 2), or because the Committee was convinced of the accuracy of the complainant’s version of the events (Outcome 3). Whatever the route followed, the violation outcome stood. The one does not exactly amount to the other, however. A procedural shift-based violation seems weaker than one reached substantively, thus with the facts at issue being positively established to have taken place.
The Committee may purposefully have avoided clarifying the basis of its decision. This could explain why, before coming to the passage quoted above, it ‘recalls’ (a word which wrongly suggests that the pronouncement which follows is a staple of its jurisprudence) that ‘it is not its task to weigh the evidence or to reassess the statements made regarding the facts or the credibility of the relevant national authorities’.Footnote 42 Such a limited interpretation of its task gestures to the no-fourth-instance principle, according to which international bodies should not (gratuitously) review domestic findings reached after possibly three sets (first instance, appeal, final instance) of proceedings. This principle, however, should be of no import when the domestic authorities have failed to (properly) investigate and establish the facts. One Committee member, Ms Gaer, vehemently objected to the words quoted above, which she described as a ‘shocking pronouncement’,Footnote 43 which went against both the Committee’s own jurisprudenceFootnote 44 and reason. To quote:
The State party [has confirmed] that its Civil Guard officers did in fact take Mr. Sonko and his companions on board their vessel and subsequently ‘released them,’ but argues that they did so ‘in an area very close to the shore,’ and did not puncture the dinghies. Thus, the State party alleges, the death of Mr. Sonko was ‘an unfortunate accident’ rather than an act of cruel, inhuman, or degrading treatment or punishment.Footnote 45
She continued:
For the Committee to reach a decision that a violation of article 16 of the Convention [prohibition of ill-treatment] occurred, it must reject the version of the facts articulated by the State party in this case. […] It is not the case that every death that occurs in custody constitutes a violation of the Convention […] In [a] case [such as this one], the Committee is called upon […] to resolve [the] tensions between the accounts offered by the author and the State party, such as whether the State’s agents left Mr. Sonko in possession of his dinghy when they expelled him from the patrol vessel […]. The Committee has apparently determined that the State’s version of the events is not credible. It is well within its power to do so and should have stated so plainly.Footnote 46
The last three lines suggest that the Committee found in fact for the complainant rather than simply considering that the State had failed to discharge its burden of proof. However, the route followed was not made clear.
This lack of transparency in reasoning is to be lamented on two important grounds. First, it risks raising the evidentiary standard which the Committee applies in future cases, since its decision suggests that it felt that the evidence before it was not sufficient for a straightforward substantive finding of violation to be reached – despite the fact that in this case its members were apparently more convinced by the version of the complainant than by that of the state. The Committee’s hesitation may influence future jurisprudence, which would implicitly, perhaps even unconsciously, expect a standard of proof higher than the balance of probabilities to be met.
Second, whether or not the dinghy had been pierced was a crucial question – not only factually but also legally. Admittedly, on paper, failing to protect life and intentionally endangering it both amount to the violation of the same state obligation (here Article 16 CAT). However, the former is an omission, the latter a positive action, so that the two are different in nature and gravity. By not having come to a determination that the dinghy must have been (presumed to have been) pierced, but having simply found that the State had failed to protect life, the Committee’s decision has less of a legal import than it otherwise would have had. This illustrates the obvious point that evidentiary reasoning impacts material obligation findings. Otherwise said, express factual findings are central to the very implementation of human rights. It is therefore essential that a fair and clear evidentiary framework be defined and applied.
2.5.2 Diory Barry v. Morocco
The second case, Diory Barry (or Diodory Barry)Footnote 47 v. Morocco, was decided in 2014, again by CAT.Footnote 48 The complainant explained he had set from Mauritania for the Canary Islands by pirogue in a perilous journey in which some thirty of his companions had died. Two weeks after departure, the forty or so remaining survivors were intercepted by the Moroccan authorities, taken to a detention camp in Morocco, kept there for roughly ten days, at which point the Moroccan gendarmerie took them to the desert in the border area separating Morocco and Mauritania. There they were told to walk to the closest Mauritanian town, approximately fifty kilometres away, each having been given a pair of plastic flip-flops, a bottle of water and a few sandwiches – no additional food or water, no protection against the sun and no blankets. Their itinerary crossed a minefield, causing the death of a person the day before the group was found, some of its members seriously wounded – including from the time they were adrift at sea. The complainant argued his expulsion had happened ‘in a flagrantly illegal manner’ with ‘no official record of his detention or expulsion’,Footnote 49 and had resulted in inhumane treatment.
The State admitted the interception of a vessel by the Moroccan Royal Navy in the area referred to by the complainant, and the detention and expulsion to Mauritania of the people found on this vessel. In a typical line of paper pushback defence, it stated that this expulsion had been executed ‘in accordance with the legislation in force’, which ‘guarantees the right of aliens to access to a lawyer, an interpreter and a doctor, to communicate with their consulate or a person of their choice and to appeal a decision to escort them to the border within 48 hours of being notified of the decision’.Footnote 50 The admitted interception, however, had not occurred on the date given by the complainant, and had concerned a larger group than the one he had described. This led CAT to observe (in a paragraph which we shall quote in its entirety, although doing so little by little rather than in one go):
The State party has partially confirmed the facts as submitted by the complainant, although confusion regarding dates and the number of migrants on board the intercepted boat means that it cannot be proven beyond doubt that the State party and the complainant are referring to the same boat and group of migrants.Footnote 51
This convoluted formulation (which resembles – without fully adopting it – the Common Law standard of ‘proven beyond reasonable doubt’) appears to want to convey that the two parties could not have been referring to the same event. Despite this, CAT took it that the State had ‘confirmed the facts as submitted by the complainant’, but without specifying which of these facts exactly.Footnote 52
The paragraph we have started to quote continues as follows:
Given that the State party has not provided any observations on the treatment of the complainant while the decision to expel him to Mauritania was being implemented,Footnote 53 due weight must be given to the complainant’s allegations.Footnote 54
With this the State is implicitly castigated for having failed to address in its submissions the way the complainant had been treated during the expulsion. This criticism does not lead CAT to apply the principle according to which a fact alleged by one party and left undisputed by the other is accepted as established, however. Instead, CAT merely derives from the absence of refutal by the State that ‘due weight must be given to the complainant’s allegations’. This formulation, routinely found in CAT’s case law when the state has failed to address the core of the complaint,Footnote 55 is regrettably ambiguous. It can mean at least three things: (1) that in the absence of refutal by the respondent state, CAT is simply to accept the complainants’ allegations (which it indeed does in some cases);Footnote 56 (2) that it has given these allegations significant weight; or (3) that it has been mindful to carefully assess which weight to grant them before finding them established.
In Barry, the ‘due weight’ formulation cannot be used in the first sense, since CAT carries on its reasoning by stating:
The Committee also refers to its concluding observations,Footnote 57 adopted following consideration of the fourth periodic report submitted by the State party, in which it expressed its concern regarding information received that, in practice, ‘undocumented migrants [had] been escorted to the border or otherwise expelled in violation of Moroccan law without having been given the opportunity to exercise their rights. Several allegations have been made that hundreds of migrants have been abandoned in the desert without food or water.’ The facts reported by the complainant are therefore not isolated.Footnote 58
Such referencing of a periodic monitoring report is permitted under CAT’s Rule 3 (2), which rightly allows use of knowledge obtained outside the particular case at hand, including as to the existence of ‘a consistent pattern of gross, flagrant or mass violations of human rights’.Footnote 59 This knowledge can potentially play a crucial evidentiary role, contributing to the factual basis of a complaint being upheld, whether procedurally (Outcome 2) or substantively (Outcome 3).
The periodic report could have been sufficient evidence for establishing that the first context condition towards operating the shift was fulfilled. Barry, however, never mentions the burden of proof or its shift. In this particular case, we suggest it is because the decision actually constitutes a substantive finding, as confirmed by the last quoted sentence referring to ‘the facts reported by the complainant’,Footnote 60 a choice of words which puts at a distance the idea that the complainant would have offered mere allegations that were upheld only because they were not rebutted by the State. If deliberately chosen (which of course we cannot know for sure), the word ‘reported’ suggests that CAT had no doubts about the facts of the case. Putting it in a positive light, the Committee would have felt certain about them.
Whilst we approve of a substantive finding having been made in this case, we regret the appearance earlier in the reasoning of the ‘due weight’ phrasing, which may suggest a procedural upholding (Outcome 2), therefore muddying the waters. If a procedural upholding, then the decision should have expressly stated that the respondent state had failed to rebut the presumption of responsibility which had arisen against it. If a substantive upholding, the decision should have spelt out that its factual findings rested on the substantive strength of the evidentiary material before it – including their alignment with what the Committee knew about Morocco’s pushback practices. Due to the lack of clarity in the reasoning, we propose to classify Barry as an ‘Outcome 5’, which ambiguously sits between Outcomes 2 and 3, as per Figure 2.1.
Adding a fifth outcome (ambiguous upholding) to our matrix.
Figure 2.1 Long description
The bottom half of the circle indicates cases in which the facts remain in dispute by the end of the adjudicatory process. The top half indicates cases in which the facts are resolved. The left side of the circle indicates cases in which the factual basis of the complainant is rejected by the adjudicator by the end of the adjudicatory process. The right side indicates those cases in which the complainant's factual basis is accepted. The outcomes start from the third quadrant and continue anti-clockwise with outcome 1 in the third quadrant and outcome 4 in the second quadrant. Outcome 1. Dispute remains or rejected. Outcome 2. Dispute remains or upheld. Outcome 3. Dispute resolved or upheld. Outcome 4. Dispute resolved or rejected. Outcome 5 appears on the right half of the circle, across the horizontal axis, between Outcomes 3 and 2, indicating that it is not clear whether the facts have been resolved or are still in dispute. In Outcome 5, the legal basis for the upholding is ambiguous.
2.5.3 D.D. v. Spain
The third case to be reviewed is D.D. v. Spain, decided by the Committee on the Rights of the Child (CRC) in 2019.Footnote 61 With support from ECCHR, D.D. complained to have been no-paper pushed back to Morocco on 2 December 2014 from the top of a border fence located on the Melilla enclave (which is Spanish territory), directly into Morocco in the custody of Moroccan officers. Even though he was 15 years old and unaccompanied, D.D. did not benefit from any process or identification in Spain. He was simply expelled, despite ill-treatment risks in Morocco.Footnote 62
The State initially disputed only his minor status and the alleged illegality of his handling.Footnote 63 Subsequently, it started to make submissions classically found in no-paper pushback cases. For example, it disputed D.D.’s identity. It stated that ‘the author ha[d] not proved that he [had taken] part in an assault on the border fence in Melilla and that there [was] no record of how he [had] entered Spain’.Footnote 64 It additionally relied on a lack of correspondence in the names that appeared on two documents D.D. produced (records of his registration in Melilla submitted to CRC to prove his presence at the border around the time of the contested pushback on the one hand; ID documents from when the Spanish State later registered him in Madrid as a minor on the basis of his Malian documentation on the other hand).Footnote 65 The State finally alleged that the complainant did not appear in the photographs he submitted, which a journalist had taken of the 2 December pushback.Footnote 66
That the state would deny at the admissibility stage that a complainant has victim status is nothing unusual in no-paper pushback litigation. Such a submission is nonetheless rare in international human rights adjudication: the facts having normally been established by their domestic courts, states do not commonly argue that an international complaint is manifestly ill-founded by disputing its facts. When a state does this, this may well in itself signal that the case deserves a full assessment, i.e., an examination that takes place at the merits stage.
At admissibility, the only thing which can be decided is whether the case is admissible, that is, whether it deserves full consideration later – nothing else. In D.D., however, the Committee proceeded to resolve the question of the disputed identity already in the part of its decision addressing admissibility. This was not the right place to do this. When a state contends that a claim is inadmissible as ill-founded on its legal substance, UNTBs routinely start by briefly addressing this point in the admissibility section, confirming that the substance of the claim is not ill-founded, before delving into a more in-depth analysis of the legal claims in the merits section. The same should happen when the facts are in dispute. They should first be assessed at admissibility and then again at the merits stage.
At admissibility, the standard of proof to be applied when assessing the facts should be prima facie. In practice, as Lisa Reinsberg explains in Chapter 3,Footnote 67 this means that the alleged facts’ plausibility (nothing more) must be verified without the adjudicator paying any attention to any state submission. At the merits stage, the applicable standard is higher, namely, more probable than not, as noted above. Not systematically going through one and then the other stage risks the two thresholds becoming somewhat amalgamated, possibly resulting in an upsliding of the prima facie standard, thus a negation of the raison d’être of the admissibility stage, and the wrong dismissal of cases as inadmissible.
Leaving aside the question of the stage of the proceedings at which factual determination should have occurred, we now turn to analysing D.D.’s evidentiary reasoning. CRC found in favour of the complainant regarding his identity, due to the State having been in the unique position to compare fingerprints but having failed to do so. For the first time in its own case law, but citing decisions by the Human Rights Committee (HRC) on secret detention, torture and enforced disappearances,Footnote 68 the Committee stated:
The Committee considers that the burden of proof cannot rest solely on the author of the communication, especially given that the author and the State party do not always have equal access to the evidence and that frequently the State party alone has access to the relevant information.Footnote 69
This formulation could have indicated that the context condition for operating the shift was fulfilled. Such an interpretation would appear supported by the references which CRC included (but which our quotation omits) to jurisprudence commonly described by legal scholars as exemplars of the shift. However, there was no shift being operated in D.D. as no presumption was identified which it would have been for the State to rebut. Instead, what the Committee operated was an explicit distribution of the burden.
This burden, CRC said, ‘[could not] rest solely’ on the complainant. The Committee weighed the arguments and evidence provided by both parties to arrive at its conclusion – a substantive upholding of the complainant’s factual allegation (Outcome 3). It is regrettable that the reasoning did not make clear that the outcome was in no way procedural. As formulated, the last quoted sentence is reminiscent of the shift, at the same time as the State’s failure to compare fingerprints is emphasised. The latter element could be mistaken to indicate that a higher level of evidence (i.e., actual knowledge that the State refrained from comparing fingerprints in its possession) than is legally necessary is required for the shift to be operated.
Spain had also disputed whether the complainant was on Spanish territory. In this regard, the Committee stated: ‘according to the author, he was arrested by Spanish security forces at the third fence of the Melilla border crossing’.Footnote 70 The italicised qualification is regrettable since, beyond the identity question, Spain had accepted that its agents had handled someone on 2 December, whom they had taken from the third Spanish border fence to Moroccan authorities. What Spain disputed was the qualification of this handling as an expulsion.Footnote 71
The complainant had advanced the existence of a risk of ill-treatment and further refoulement upon expulsion to Morocco.Footnote 72 Spain addressed this point factually, neither accepting nor rebutting it.Footnote 73 Without noting this, the Committee upheld it, neither as an undisputed fact, nor within the mechanism of the shift, but relying on its own reporting.Footnote 74 This reporting – as Hakiki knows for having been part of the team litigating the case – had been specifically submitted by the complainant, alongside other reports by the UN Special Rapporteur on Torture, the BBC, Human Rights Watch and Médecins Sans Frontières. The latter reports are mentioned in the decision in relation to the factual issue of the content of Spanish policyFootnote 75, but regrettably not in relation to the risk of ill-treatment upon expulsion. It would have been better for the Committee to acknowledge these reports’ probative value in corroborating its own reporting.
Section 2.5.1, we have observed in relation to Sonko that a vague assessment of the evidence may lead to material state obligations being left undetermined. The same phenomenon is visible in D.D. Here the Committee did not assess whether the evidence showed the existence of ‘substantial grounds for believing that there is a real risk of irreparable harm to the child’ if the child were to be expelled (Article 37 of the Convention for the Rights of the Child). The Committee restricted itself to concluding that the evidence showed that the State had not assessed whether such substantial grounds existed. A violation of procedural obligations under Article 37 was confirmed,Footnote 76 but the related substantive human right was not addressed.
In conclusion, although the Committee’s evidentiary conclusions in D.D. deserve to be broadly applauded, explicit and precise reasoning is lacking. Not only does it leave practitioners in the dark as to which evidentiary rules apply, but it also weakens human rights accountability.
2.5.4 A.B and B.D. v. Poland
The last case to be considered is A.B. and B.D. v. Poland, decided by the HRC in 2023.Footnote 77 It concerns a paper pushback, here of a Chechen family who had presented themselves over twenty times to the Polish authorities at the border crossing from Belarus to Poland, and had each time been issued with a refusal of entry.Footnote 78 The complainants eventually gained access to the Polish asylum framework: after HRC itself granted them interim measures for Poland to register their asylum claims and to refrain from removing them from the territory until its own decision on the merits of their complaint.Footnote 79
As typical of paper pushbacks, the key disputed point between the parties was whether, in their previous interactions with the Polish officers, the family had attempted to claim asylum. The complainants claimed this,Footnote 80 referring to national judgments confirming the existence of a state practice of denying access to asylum procedures, findings by various UNTBs and other human rights institutions and organisations as to systematic pushbacks taking place at the Polish borders, and governmental public declarations confirming the same.Footnote 81 In contrast, the State submitted that the complainants had been interviewed as to their reasons to seek entry into Poland and had only ever expressed economic or personal reasons, as per ‘reports of the Border Guard officers’.Footnote 82
The first remarkable feature of this decision holds in its opening. Usually, a decision starts by simply recording the identity of the complainant, the legal provisions relied upon in the complaint, and the name of the representative(s). In this case, however, the decision begins as follows:
The authors fled the Russian Federation and arrived in Belarus in January 2017, with the intention of applying for asylum in Poland at the border between Belarus and Poland, in Terespol. The authors made more than 20 attempts, in the period from January to August 2017, to request asylum at the border. On each of those occasions, however, their claim for asylum was not recognized as such by Border Guards with the result that the claim was not passed on to the competent authorities for consideration. As a result of their not having valid travel documents and not being acknowledged as having claimed asylum, the authors were issued decisions by which they were denied entry into Poland and were immediately expelled and sent back to Belarus.Footnote 83
What is extraordinary in this quotation is that the Committee adopts the complainant’s version of the events from the very start.
Whilst there is no explicit explanation as to why these events were not reported in the section of the decision entitled ‘Facts as presented by the author’, the seasoned observer can later guess what led the Committee to proceed in the way it did. The next paragraph records that the Committee had examined and granted an application for interim measures upon submission of the complaint on 24 and 25 August 2017.Footnote 84 The section entitled ‘Facts as presented by the authors’ subsequently reports that the complainants were again denied access to asylum at the border post on 29 August 2017. (This would normally have been mentioned in the introductory paragraph summarising the procedure). In the section dealing with admissibility, the Committee notes that whether the complainants had claimed asylum was disputed by the parties,Footnote 85 but resolves it simply by stating: ‘the authors have sufficiently substantiated their claims that they presented information to authorities at the Terespol border checkpoint’.Footnote 86 In the section devoted to the merits, the Committee yet again notes the State had brought up this issue.Footnote 87 This time it does not address it itself, presumably because it considers it was resolved at admissibility.
Given the State’s non-compliance with the interim measures, followed by its shameless denial that the complainants had ever wished to claim asylum, against the Committee’s own finding that they had had such a wish, as confirmed by its very request to the State that they be given access to asylum procedures, the Committee could not but be entirely confident about the facts. A substantive upholding of the factual basis of the complaint was therefore the right outcome in this case. However, the Committee should have expressly explained this, as well as recorded that the standards of proof applicable at the admissibility and merits stages differ.
2.5.5 Conclusion
The analysis offered in this section can be summarised as follows: UNTBs are to be commended for having, to this day, seen through the state denials and for having invariably accepted the factual allegations of the complainants in pushback cases, despite the evidentiary challenges these cases raise; however, their evidentiary reasoning would gain in being sharper. What is needed, then, is not only clarity regarding the identification of applicable principles but also appreciation of the negative impact of lack of clarity. Section 2.6 starts this conceptual work by highlighting the pitfalls that arise when the evidentiary framework is loose.
2.6 The Risks of a Vague Evidentiary Reasoning
Having demonstrated in Section 2.5 that UNTB jurisprudence suffers from a lack of evidentiary clarity,Footnote 88 this section seeks to highlight the risks which this weakness generates for human rights accountability by analysing in further depth two ‘clouding’ elements. We argue that the absence of specification of the standard of proof applied by the adjudicator, commonly lamented in the literature, risks producing an inadvertent raising of the standard of proof to the detriment of the complainant. The second element we have selected for discussion relates to our recommendation that the either procedural or substantive manner in which the factual basis of a complaint is upheld be distinguished. Confusing these two routes, we argue, risks obliterating the obligation for the adjudicator to uphold the complainant’s factual allegations when the circumstances for reversing the burden of proof are present.
Table 2.3 helps visualise these two risks. Its vertical axis lists the evidentiary standards of proof that can be reached when seeking to substantiate a complaint’s factual basis. The horizontal axis lists the three outcomes discussed above.Footnote 89 The table’s top three circled areas indicate where the two axes normatively intersect, that is, which standard of proof is applied that triggers a particular outcome.
Matching evidentiary thresholds and outcomes (merits stage)
Table 2.3 Long description
If the Standard of Proof reached is Prima facie, the outcome is Outcome 1 except if the burden of proof is to be shifted onto the state under Procedural rejection, or Outcome 2's linkage condition S o P near-universally recognized under Procedural shift-based upholding.
If the Standard of Proof reached is More probable than not, the outcome is context condition Roberts under Procedural shift-based upholding, or Outcome 3 S o P widely recognized, under Substantive evidence-based upholding.
If the Standard of Proof reached is Clear and compelling, the outcome is context condition Dembour and Hakiki under Procedural shift-based upholding, or Outcome 3 even more warranted under Substantive evidence-based upholding than the one above.
If the Standard of Proof reached is Beyond reasonable doubt B R D, the outcome is Outcome 3 even more warranted under Substantive evidence-based upholding than the two above.
If the Standard of Proof reached is Absolutely certain, the outcome is Outcome 3 infinitely more warranted under Substantive evidence-based upholding than the three above.
Outcome 1 for Prima facie is highlighted in a bright shade. Outcome 2 for Prima facie and Outcome 3 for more probable than not are highlighted in a dark shade. Outcome 3 for clear and compelling, B R D, and absolutely certain are highlighted in lighter shade.
The ‘more-probable-than-not’ standard, also known as preponderance of evidence, is the standard of proof generally accepted to be of application in international human rights adjudication at the merits stage.Footnote 90 By definition this standard is respected when higher standards are met, as Table 2.3’s last column’s three bottom light grey boxes capture.
These same three boxes indirectly suggest how the standard of proof imposed on the complainant could inadvertently be raised. For the sake of this discussion, let us assume that a couple of pushback cases were decided by a committee on unusually strong evidence, possibly enhanced by technological operations (as in Parvin A and I.M. discussed above). If this is done without appreciating that the submitted evidence was stronger than legally required, committee members may come to expect the same level of evidence to be reached in subsequent cases. In the absence of clearly set and understood standards of proof, previous adjudicating experience risks becoming the adjudicator’s point of reference. Academic studies have observed that the prima facie standard of proof commonly gets distorted in human rights adjudication in a way that disadvantages the complainant victim. The Strasbourg BRD standard is similarly affected.Footnote 91 We speculate that an upward slippage could also affect the ‘more probable than not’ standard.
Table 2.3’s two dark--shaded encircled boxes help visualise the risk of an incorrect application of the shift. They show that a violation finding can only be reached procedurally – through the operation of the shift – if the evidence linking the complainant’s complaint to the problematic human rights contextual situation is no stronger than prima facie. Additionally, they show that when there is plausible evidence of this link (and the underlying problematic context has been proven to a higher standard, and the state has failed to rebut the presumption which has arisen against it), then the adjudicator must find in favour of the complainant factually. If this were not an automatic outcome, there would be no point in envisaging the shift as a corrective mechanism in the evidentiary framework when states for example suppress evidence of their violating practices. If the standard which is met is the balance of probabilities, then the case is a standard one, represented in Table 2.3 by the darkest-shaded box (substantive upholding).
2.7 Conclusion
It is crucial that human rights adjudication takes place in evidentiary normative fairness, clarity and consistency. Three points in particular have emerged for special attention in the course of our analysis of the UNTB pushback case law. Firstly, in international human rights adjudication, disputed factual allegations must be examined first at admissibility – under the prima facie standard of proof, and again at the merits stage – under the preponderance of evidence standard. Secondly, whilst a complainant will not suddenly stop their submissions when they think they have achieved the minimum standard of proof legally required, the adjudicator must, for its part, recognise when the submitted evidence has gone beyond the set minimum legal standard so as to avoid the understanding of the requirements of relatively low standards being skewed and slipping upwards. Thirdly, for the shifting of the burden of proof to remain a meaningful corrective to the application of the normal evidentiary principles when the state attempts to escape human rights accountability, its operation must take place in strict adherence to the conditions and mechanisms of its application. This requires, amongst other elements, that the adjudicator identify the cases where the shift is not appropriate, due to a substantive upholding of the factual basis of the complaint actually being indicated.
Open access