From Invisibility to Protection: Sexual Violence as Persecution and the Case for a CEDAW-Informed Refugee Framework

Introduction

Seventy-five years after the Convention Relating to the Status of Refugees¹, sexual violence for various reasons is displacing populations in South Asia, the Middle East and Sub-Saharan Africa to an extent more than what the Convention’s original drafters could have predicted. The Convention’s definitional architecture was never developed with the gendered aspects of persecution in mind, and was created in the wake of the Second World War, but largely based on the experiences of European male political dissidents. Over the last 70 years, the Convention has been developed to cover claims based on sexual violence through the guidance of UNHCR², through domestic jurisprudence and through scholarly writing³. But it has been a gradual process, lacking in uniformity in boundaries, doctrine, and even in timing. However, a protection architecture remains in place that makes it structurally challenging to recognise survivors’ experiences of sexual violence as persecution within the Convention’s terms, and that still appears to be in place in Asia and the Global South, where survivors are seeking protection.

This gap in protection as identified in this paper is a structural gap and not an incidental one. It is situated, first, in the inconsistent and contested use of the term of the “particular social group” as the main surrogate basis on which to base a claim to international protection, for which UNHCR often has no guidelines; and second, in the frequent failure to apply UNHCR’s Guidelines on International Protection Nos. 1 and 2⁴, there is the absence of persuasive, but non-binding, statements on the Convention, namely, General Recommendations 32⁵ and 35⁶, and third, the failure to consistently apply the Convention in RSD practice, particularly when reading it in conjunction with these two documents that provide the most elaborated international articulation of gender-based violence as a form of discrimination and harm. This protection system is a system where, while sexual violence is granted recognition in principle, it is not structurally recognised in practice.

This paper is based on three related questions. There is first the question of why the existing apparatus of interpretation of the Convention of 1951 remains insufficient to define sexual violence as persecution, in light of the Convention’s treatment of “particular social group”. Second, does CEDAW and in particular General Recommendations 32 and 35 have the capacity not only to provide an additional interpretive tool but to also serve as a re-imagined framework for the understanding of persecution, well-founded fear and State protection? Third, how does this reconceptualization call for a change in institutions, notably in India, which has neither signed nor acceded to the 1951 Convention Relating to the Status of Refugees or the 1967 Protocol, means that complementary protection is the main recourse available to survivors? The paper thus aims to highlight the doctrinal and structural shortcomings in the Convention’s gender neutral drafting when considering claims for sexual violence, to conceive of CEDAW-informed reinterpretation as a constitutive rather than an ancillary interpretative tool for analysis of persecution, and to bring this reinterpretation to life in the context of concrete institutional and procedural changes, particularly in the context of RSD practice in South Asia and India.

India has neither signed nor acceded to the 1951 Refugee Convention or the 1967 Protocol. Consequently, refugee protection in India operates through constitutional guarantees, executive discretion, and the mandate Refugee Status Determination procedures of the United Nations High Commissioner for Refugees, making complementary protection the principal avenue available to survivors of sexual violence.

From a methodological perspective, the paper is doctrinal and normative, while keeping an eye on the Third World Approaches to International Law (TWAIL) sensibility which is sensitive towards the international refugee protection regime’s structural position of the Global South. It is based on the language of the treaties, the principles of treaty interpretation as set out in the Vienna Convention on the Law of Treaties, in particular under Article 31(3)(c)⁷, the UNHCR guidance, comparative asylum case law, and the General Recommendations of the CEDAW, along with an applied case study of Indian State practice.

The main thrust of the article is that protection gap for survivors of sexual violence in the Convention could be addressed without any formal amendment, rather by reconceptualizing the core categories of persecution, well-founded fear and State protection, that are defined in the Convention in the light of the CEDAW. It argues that the change in paradigm from an interpretive to a constitutive scaffolding of CEDAW, is necessary but not sufficient and, if the Convention is to become a reality for survivors in the Global South, must go hand in hand with targeted institutional reform.

Sexual Violence as Persecution under International Refugee Law

The refugee definition contained in the 1951 Convention is based on the fundamental notion of persecution, but the Convention does not define the term⁸. This is not by chance, but planned and left intentionally open to interpretation and has also permitted the idea to develop differently in various jurisdictions. The dominant perspective on persecution, which is closely associated with the definition of “sustained or systemic violation of basic human rights,” sees it as a process, reflecting the nature of international human rights law as a body of law, rather than as a set of harms. This human-rights-based approach, at the heart of Hathaway’s line of thinking that persecution is a denial of basic human dignity⁹, was only partially and unevenly taken up in case law and doctrine. This imbalance, as will be shown in later parts of this paper, is not an incidental or a minor defect in the Convention, but rather a structural problem in the Convention, and it is this disconnect between what is conceptually possible and what is doctrinally realised that drives this reinterpretation of the Convention in the light of CEDAW.

The idea of sexual violence as a particular type of international legal damage has taken the greatest form of development outside of refugee law. The ad hoc international criminal tribunals added doctrinal depth to the ideas of rape as a grave breach and as a war crime, which were already part of the language of international humanitarian law contained in the Additional Protocols of the Geneva Conventions¹⁰. The International Criminal Tribunal for Rwanda in Prosecutor v Akayesu¹¹ is crucial in this respect, as it found that rape and other forms of sexual violence could be part of the genocide, where committed with the necessary intention to destroy a protected group; the ICTY’s judgment in Prosecutor v Kunarac¹² also held that rape and sexual violence could be considered a form of torture and enslavement under customary international law. Since then the definition of sexual violence has been incorporated into the body of international criminal law, as a crime against humanity and a war crime under the Rome Statute of the International Criminal Court¹³. International human rights bodies, at the United Nations and at the regional level (as in the Committee against Torture), have expanded the definition of rape from the criminal law to the larger human rights framework and recognized, in broad terms, that rape is a form of torture or inhuman treatment committed by the State when perpetrated by State agents or with the acquiescence of the State¹⁴. This intersection of international humanitarian law, international criminal law and international human rights law gave refugee law a normative foundation that it then lacked, an authoritative, cross-regime endorsement that sexual violence is serious harm that can lead to international legal obligations, outside of the context in which it subsequently became a refugee legal cause of action.

It is important, therefore, to note that this normativisation occurred primarily in absence of reference to the gendered structural conditions of discrimination, unequal power relations and systemic subordination which will be surfaced by CEDAW’s General Recommendations 32 and 35 as a root cause of gender-based violence, not as a trigger for it^15,16. To the extent that the international legal community recognized rape as torture, genocide, or a crime against humanity, this did little to provide an understanding of why rape disproportionately affects women, or of how individual acts of rape are connected to the systemic discrimination that CEDAW considers to be the underlying cause. This lack, which continues to be unfilled in refugee jurisprudence, is the space which Section IV of this paper contends that is uniquely filled by CEDAW.

Refugee jurisprudence built on this normative undercurrent and accepted sexual violence as persecution in various aspects of harm, albeit with some differences in doctrinal certainty. Physical hurt was easiest to accept, as it was visible and because it was very similar to torture and cruel treatment which already was considered to be paradigmatic persecutory acts that did not need to be explained for long. Psychological harm was more contentious: adjudicators had to be convinced that trauma, humiliation and social ostracism that followed sexual violence could in and of itself meet the severity requirement for persecution, especially in societies where the survivors are being rejected, or subjected to additional persecution, by their own communities and not just the initial perpetrators, a compounded harm not adequately recognised by the existing doctrine of persecution, which is centred on a single identifiable persecution act¹⁷.

Adding to the difficulty was the issue of State complicity. In cases where State agents (such as security personnel, police or military personnel) were directly responsible for committing sexual violence, the sexual violence was characterised as persecution attributable to the State, and the attribution was therefore made in adherence to the established principles of State responsibility, which required little more than the application of the existing attribution doctrine to a new set of facts. Questions of whether State was unable or unwilling to provide protection, by contrast, have been far more persistent as evidenced and doctrinal concerns, especially in the type of RSD system that is most likely to face such questions, given the evidence typically available to a single claimant, to assess on the evidence whether protection against gender-based violence is practically adequate in the country of origin. It is in this protection inquiry, that is not undertaken by the State, that the due diligence standard of CEDAW is best suited to look beyond an individualised and frequently hypothetical analysis of whether protection was due in one particular case to a structural assessment of the State’s demonstrated pattern of response.

Sexual violence as a form of persecution has, on this uneven basis, evolved in an uneven manner, particularly in a small group of common law jurisdictions. The House of Lords’ judgment in Islam and Shah v Secretary of State for the Home Department¹⁸ was groundbreaking for the idea of a social group of women who have suffered domestic violence and are thus at risk of false accusation, thus providing an opening for the interpretation of women as a social group for gender-based claims in general, rather than giving a final answer on the matter. The U.S. Board of Immigration Appeals in Matter of Kasinga¹⁹ had come to a similar conclusion recognizing the threat of female genital mutilation as persecution for social group status, an opinion which, more than 20 years later, continues to be referenced in U.S. asylum litigation. Building on the Supreme Court of Canada’s seminal decision in Canada (Attorney General) v Ward,²⁰ Canadian jurisprudence further unified the doctrine of social-group membership based on gender by articulating the categories of analysis, including those involving groups based on “innate or unchangeable” characteristics, that guided the assessment of gender-based claims.

All these choices are real doctrinal progress and further expand the Convention’s protections to those who the original Convention’s authors could not have envisioned. They are equally evidences of the very imbalance that this paper points to as a deeper structural problem in RSD practice, one that does not simply involve the attitudes and prejudices of individual adjudicators but the Convention itself, and more particularly the instability of the notion of ‘particular social group’ as the central feature through which the Convention’s architecture has been made to serve claims it was never intended to meet. This paper now focuses on this structural limitation, and the case for the possibility that it can be overcome by CEDAW-informed reinterpretation instead of further incremental adjustment.

Structural Limitations of the 1951 Refugee Convention

The definitional architecture of the 1951 Convention reflects a certain historical moment, and is formulated in the context of the Second World War, the experiences of European political dissidents and stateless persons. The five grounds listed are “race,” “religion,” “nationality,” “membership of a particular social group” and “political opinion,”²¹ and capture the paradigmatic forms of persecution known at the time state action against an individual on the basis of their identity or their beliefs. Gender was not listed as an enumerated ground and there was no continued discussion of the nature of harm suffered by women that is disproportionate or specific to them. This was no simple mistake that can be made up through practice, it is a conceptual weakness in the Convention that has defined, and limited, how sexual violence claims are adjudicated today.

The grounds listed in the Convention have been used over the years to support gender-based claims, and sometimes with mixed results. Other than when it involves ethnic conflict, e.g. sexual violence as a tool for ethnic cleansing, race/nationality has little relevance to sexual violence claims. For example, religion has sometimes been used as a basis on which sexual violence has been committed to enforce religious codes of conduct, or to punish perceived transgression of religious norms about the behaviour of women. While in such cases the harm may be said to be religiously motivated, the actual basis for the sexual violence is usually considered to be gender-based rather than religious. Political opinion, on the other hand, has been more flexible, as the doctrine of “imputed political opinion” has been adopted in decisions (including the Supreme Court of Canada’s ruling in Canada (Attorney General) v Ward)²² where persecutors may infer political opinion from a victim’s actions even if he or she does not harbour a specific opinion. Despite this, political opinion analysis is not always compatible with sexual violence allegations made because of a more political than sexual motivation, namely to assert gendered social control, the subjugation of women to men, or to impose communal norms on their sexuality and behaviour.

It is therefore the residual term of ‘particular social group’ (PSG) that has thus far carried the main burden of the doctrinal argument in support of claims on the basis of gender, precisely because of the openness of its meaning that allows it to serve as a proxy for harms that were not considered by the drafters of the Convention. In the United States, the interpretive basis for PSG analysis started with Matter of Acosta²³, which adopted a definition of a particular social group as having a number of members who share a common immutable characteristic which is either natural or of such fundamental importance to identity that it can never be altered. This “immutability” approach was followed by the “social perception” approach of “Australian jurisprudence” which asks the question, rather, does the surrounding society perceive the group as a distinct group?²⁴ Capturing either of these two tests, UNHCR’s Guidelines on International Protection No. 2 offer a disjunctive formulation, treating a group as a particular social group when it meets either the protected-characteristic test or the social-perception test²⁵. But the reconciliation has failed to bring about the doctrinal stability the reformers had hoped for: Adjudicators in various jurisdictions use the two tests in inconsistent ways, sometimes both tests, in the same jurisdictions, and and the results are therefore unpredictable for claims of a similar functional nature.

This is exacerbated by an additional, more structurally problematic issue: the circularity in PSG analysis, that is, the definition of the “particular social group” in terms of the persecution. Decision-makers have, at different times, described the group as “women who have suffered domestic violence” or “women who fear sexual violence”, which may result in the requirement of a group of women becoming re-described as the harm suffered. The House of Lords’ judgement in Islam and Shah²⁶ above made it more difficult to do this by attaching the social group to a wider social characteristic, namely the status of women in Pakistan, as a social group, under systemic discrimination and lacking adequate legal protection, as opposed to the experience of violence alone, but this more rigorous approach in other jurisdictions has not been consistently followed. The result is a body of PSG doctrine that remains unsettled seventy years after the drafting of the document, and it remains unsettled at precisely the threshold question on which gender-based claims most often turn: whether the claimant belongs to a cognisable group at all.

In addition to PSG’s internal instability, there is another obvious evidentiary challenge that is unique to claims of sexual violence: RSD systems that rely on an assessment of the credibility of testimony from people who have suffered persecution for political or religious reasons are not necessarily well-suited to the evidentiary and psychological realities of sexual violence disclosure²⁷. Such inconsistencies are well documented in trauma-informed interviewing literature, as are patterns of survivors’ delayed disclosure or refusal to disclose sexual violence, and these are typically handled as evidence of fabrication and not as a consequence of trauma in the adversarial or quasi-adversarial RSD procedures²⁸.

All of these constraints, the ground-neutral approach of the Convention, the fluidity and lack of stability of PSG as the main surrogate category, the potential for circular reasoning in defining groups, and the evidentiary disconnection of RSD procedure and disclosure of trauma, show that the protection gap addressed in this paper will not be filled by doctrinal tinkering. But rather than a further refinement of PSG analysis, what is needed is a new point of departure altogether: one that can provide the substantive content which the Convention’s gender-neutral categories literally do not have. It is to this alternative interpretive framework, which is based on the CEDAW, that the paper now moves.

CEDAW as a Constitutive Interpretive Framework

There is no parallelism between international refugee and international human rights law, but rather there is a systemic integration, if there is any. The principle of interpreting the treaty “in the light of” any “rules of international law applicable in the relations between the parties” given in Article 31(3)(c) of the Vienna Convention on the Law of Treaties (1969)²⁹ authorises, indeed perhaps demands, the 1951 Convention to be read in tandem with other human rights instruments that have been developed since 1950, rather than being read in isolation at the horizon of 1950. It is a position not being discussed here for the first time: UNHCR’s own guidance has always included a human rights-based approach to refugee status determination and the “sustained or systemic violation of basic human rights” criterion for persecution is based upon the assumption that engagement with the developing body of international human rights law is necessary to understand what constitutes a basic human right and what constitutes a sustained violation³⁰. In other words, the Convention itself is not self-contained, but is derived from, and is subject to, the human rights regime which is the standard by which persecution is judged. This paper asks not whether it is permitted to cross-reference with other documents, but which human rights instrument to use for providing the “content” for gender-based persecution claims and how much “interpretive authority” can be given to that instrument.

It is precisely because, of all the human rights treaties, CEDAW is the most developed treaty in this regard (the object and purpose of which is the elimination of discrimination against women in all its forms), and the treaty body which has developed authoritative and detailed guidance on gender-based violence as a manifestation of that discrimination³¹. The General Recommendation No. 19, and the consolidation and expansion of this in General Recommendation No. 35 by the Committee on the Elimination of Discrimination against Women, defines gender-based violence against women as discrimination as defined in Article 1 of CEDAW, whether the perpetrator is a State agent or a private individual. The due diligence obligation is expanded upon in General Recommendation No. 35^32,33 which requires that States take appropriate measures to prevent, investigate and punish violence by non-State actors as well as structural and institutional conditions that facilitate the continuation of violence with impunity. In the case of General Recommendation No. 32, addressing the gender-related dimensions of refugee status, statelessness and asylum more specifically³⁴, the Committee makes demands upon States parties to ensure that the RSD procedure, evidentiary standards and credibility assessments take a gender-related perspective; and explicitly ties the application of the due diligence framework in CEDAW to the interpretation of “persecution” and “particular social group” in the 1951 Convention. It is in this sense the most straightforward textual connection between the two regimes: General Recommendation No. 32, in spite of its explicit reference to refugees, is the one that is most underutilized in RSD practice.

Under CEDAW’s due diligence framework, some of the key interpretive categories in the Convention are significantly re-orientated. First, persecution is not judged by how visible and serious the physical abuse is, but rather by the structural conditions of discrimination which systematically make women susceptible to violence and systematically inaccessible to protection and/or redress. Second, well-founded fear is re-oriented by CEDAW’s “due diligence” standard to focus not on who or why the perpetrator is, but on the State’s demonstrated pattern of failure to prevent, investigate, or punish gender-based violence, and a pattern that CEDAW jurisprudence considers to be itself probative of the risk a claimant may face on return³⁵. Third, state protection is rethought from a formal examination of the existence of protective laws to a substantive examination of whether those laws are enforceable without a discriminatory application, a shift that is explicitly noted in CEDAW’s due diligence doctrine and is neglected in PSG jurisprudence, which does not focus primarily on group protection. Fourthly, non-State persecution is therefore re-framed: the refugee law requirement for proof of a discrete failure of protection in the individual case would not be met by a non-State response lacking in such a clear pattern, but instead a structural approach to the protection liability of States would allow for the use of documented systemic patterns of State inaction towards gender-based violence to establish that the State was unwilling or unable to provide protection. The recognition in General Recommendation No. 35 of the impact of trauma on disclosure and testimony in CEDAW provides a human rights foundation—not just a psychological or evidentiary one—for refugee adjudicators to consider when assessing delayed disclosure and/or inconsistencies as evidence of sexual violence, not a fatal flaw.

The cumulative impact of this reinterpretation is to provide refugee law with an account of the nature of gender-based discrimination and violence that it lacks in an intrinsic sense, one built by the treaty body, the one that most clearly has an institutional mandate and interpretive authority, to articulate it. This is why the paper stresses the need to interpret CEDAW as in addition to, and not instead of, its constitutive nature. A constitutional role would consider CEDAW’s General Recommendations to provide substantive content to persecution, well-founded fear and State protection, and therefore, RSD analysis would be incomplete, and perhaps even incorrect, where it would occur without reference to CEDAW’s due diligence framework. This is important because invoking this as a condition or part of the condition of being protected is not required and is not foreseen by the rules as a matter of law, whereas invoking it is a duty under Article 31(3)(c) of the CEDAW, which will be subject of review as a matter of law if not done. It is this normative shift – from CEDAW as available authority to CEDAW as a required interpretive lens – that represents the paper’s main contribution and against which the paper now measures its application in the context of refugee status determination practice in South Asia in particular, and India in particular.

Refugee Status Determination and the Global South: The Indian Perspective

The interpretive framework of Section IV will have to be implemented in the refugee process, and doctrinal reform will also be limited to the academic realm if it isn’t carried out at the procedural level of refugee Status Determination. In general RSD procedures build an onus on the claimant to establish that he or she is being persecuted, which is generally understood as a burden of proof that is specific to the unique evidentiary challenges which arise in persecution cases, as reflected in the formulating of UNHCR’s RSD criteria (which require the claimant to establish a “credible account” and “the benefit of the doubt” where there is no documentary evidence)³⁶. In practice, however, this standard is applied during credibility evaluations and interview techniques that were not created with disclosure of sexual violence in mind and, therefore, result at the procedural level in a similar mismatch of evidence, as revealed in Section III. UNHCR’s own procedural standards recommend the use of gender sensitive interviewing (same-sex interviewers and interpreters, private interview settings, and interviewers trained in trauma-informed interviewing), but there is a lack of consistency in applying it, especially in countries where there is a shortage of resources or institutional resistance to specialized staffing³⁷.

This is further exacerbated by the lack of national refugee legislation in many States in the Global South, resulting in refugee status determination being in the hands of UNHCR field offices rather than in a national legal system with procedural safeguards and rights of appeal. This results in marked disparities in the outcomes of claims that are not necessarily due to the merits of the individual claims, but rather to the institutional capacity, training and procedures of the specific UNHCR office or, where such exists, the domestic authority concerned, and which is often reflected in the inconsistent recognition of GBC claims that has been widely documented in comparative studies of RSD practice in South and Southeast Asia countries, where there is no harmonisation of regional standards³⁸.

India is a case in point in this regard. There is no refugee-specific legislation in India and the Foreigners Act 1946³⁹ and the Registration of Foreigners Act 1939⁴⁰ do not distinguish between refugees who flee the persecution and other types of foreign nationals and give broad discretionary powers to the executive on entry, stay and removal. In this legislative void, protection for refugees has been largely achieved through constitutional adjudication and the mandate RSD function carried out outside of, but alongside, domestic law, with little to no collaboration or integration. The Supreme Court’s judgment in National Human Rights Commission v State of Arunachal Pradesh⁴¹ has broadened the protection under Article 21 of the Indian Constitution, which guarantees the right to life and personal liberty, to include protection against arbitrary deportation, even for non-citizens who were threatened with expulsion from the state. In Louis De Raedt v Union of India⁴², the Court had earlier laid down that foreigners have a right to be protected by Article 21 of the Constitution which includes right to a fair procedure before deportation, but not also a substantive right against deportation. The Gujarat High Court’s order in Ktaer Abbas Habib Al Qutaifi v Union of India⁴³ is even more interesting as it incorporated a fundamental substantive right of refugee law into Indian constitutional law, even though India was not a signatory to the “principle”. All these decisions together create a jurisprudential framework for complementary protection in India, albeit one that is disjointed and relies on case-by-case adjudications, without any of the structures, such as an established RSD procedure with uniform RSD criteria, appeal, or gender sensitive interviewing protocols, that a dedicated RSD system would offer.

While the government continues to lack RSD infrastructure, in this vacuum UNHCR’s New Delhi office conducts mandate refugee status determination for asylum seekers using UNHCR’s own procedural standards and criteria⁴⁴. As in Section III and above, the application of a CEDAW-informed understanding of persecution and the trauma sensitivity of the interviewing procedures are key interpretative and procedural factors that contribute to access to protection for survivors of sexual violence in India; and the scope of the definition of “particular social group” is another that may affect access. Valuable as a backstop it is, constitutional protection under Article 21⁴⁵ does not provide for anticipatory, standardised, procedural protections that a codified RSD system would provide; it does not directly engage with CEDAW’s substantive framework for the determination of gender-based persecution, and the development of Indian constitutional law in this field has occurred around the language of Article 21, not through any explicit engagement with CEDAW’s General Recommendations, despite India’s status as a State party to CEDAW itself⁴⁶.

The institutional gap can have far-reaching repercussions for those who have experienced sexual violence. When RSD or constitutional protection is lacking, the risk of a return to the conditions of harm that the Convention’s protection regime aims to address is a very real one of refoulement. Asylum seekers are often subjected to administrative detention under the Foreigners Act⁴⁷ as a consequence of the uncertainty while constitutional proceedings are pending, or while they are waiting for RSD processing, which can, for female survivors of sexual violence, be especially traumatic as it may be a repetition or exacerbation of previous trauma. The lack of gender-specific, trauma-informed procedures on all stages of this process may lead to secondary victimisation (re-victimisation) of trauma survivors due to repeated, non-gender-sensitive interviews, raising questions about the disclosure of their experiences, and lack of institutional awareness of the gendered nature of these disclosures. These are not only administrative inefficiencies but the real, human cost of the doctrinal and structural problems discussed in this paper and are a reminder of the need for the CEDAW-informed reinterpretation advanced in this paper to be put into practice through the institutional reforms which are the subject of the next section.

Towards a CEDAW-Informed Refugee Framework

The analysis in the previous sections creates both a doctrinal argument for the need for CEDAW-informed reinterpretation of the 1951 Convention and a very pressing need for such reinterpretation, in the context of the RSD systems in the Global South and survivors of sexual violence. This last analytical section is devoted to this reinterpretation and is translated into three specific institutional reforms before directly considering the question posed in this paper’s abstract: can interpretive innovation fill the protection gap or does institutional reform need to be implemented to achieve interpretive innovation?

The first suggested reform is that the content of the interpretation is formalised in the shape of a UNHCR led Protocol on Gender-Based Persecution (GBP), which brings together the guidance that is spread across the Guidelines on International Protection Nos. 1 and 2^48,49 and General Recommendations 32 and 35^50,51 of the CEDAW into one authoritative ruling specifically directed at RSD adjudicators and caseworkers. Such a Protocol would incorporate the present structure of the constitutive framework in Section IV, and would include the following, as operational definitions that RSD systems now lack, namely: an explicit definition of persecution as a structural phenomenon and not episodic; an operational standard for assessing State protection based on due diligence and not existence; and an explicit instruction and obligation that trauma-related disclosure patterns be taken into account in assessing credibility and not treated as a pointer to fabrication.

Despite being non-binding, UNHCR’s Guidelines on International Protection No. 2 helped to at least partially reconcile divergent approaches to PSG analysis by providing a shared glossary of language and a disjunctive test which could account for different domestic doctrinal traditions⁵². A Protocol “specific to” gender-based persecution, and explicitly based on the General Recommendations in the Convention, would fill a similar interpretive void, without the political and procedural stumbling block inherent in formal Convention amendment, which this paper has always aimed to evade. Importantly, a Protocol would also solve a coordination problem, now shared by caseworkers and domestic adjudicators, of having to research and synthesize material found in at least three documents, the two Guidelines and two General Recommendations, before making a decision—a burden that would be greatly diminished with a single consolidated Protocol, as noted in Section V, which identified a lack of coordination among the relevant instruments as a significant source of inconsistent outcomes across Global South RSD systems.

The second reform is to incorporate the content of CEDAW’s General Recommendations as a mandatory part of analysis and review in RSD procedures as opposed to relying on the discretion of the adjudicator, which had been criticised in Section IV. Operationally, this means that General Recommendations 32 and 35 should be integral components of a national RSD authority’s or UNHCR field offices’ RSD training programme, not provided as ‘optional extras’ to be used at the discretion of individual caseworkers, including UNHCR New Delhi. It would also have implications for judicial and administrative evaluation: If CEDAW engagement is to be required as part of RSD analysis, then its omission in a specific determination would be subject to review as an error of the law rather than as an omission that is not subject to challenge as a matter of discretion. The reform is of special importance for States which have ratified CEDAW but have not acceded to the 1951 Convention, as is the case in India. For those States, the imperative to integrate CEDAW is not to join the instrument which they have, for seventy-five years, refused to join, but to give a concrete and meaningful application of treaty obligations that India has already assumed, and thus to turn CEDAW into a substantive source of refugee protection content embedded in existing RSD and mandate-processing practice⁵³. The reform is doctrinally conservative in this regard, albeit transformative in its practice, requiring States to fulfil existing obligations they already have, without taking on further international commitments.

The third reform will be to ensure that survivors of sexual violence can benefit from complementary protection and non-refoulement guarantees under case-specific constitutional cases, rather than in the form of a codified domestic law, based on the jurisprudence developed in Indian constitutional cases, such as NHRC v State of Arunachal Pradesh⁵⁴ and Ktaer Abbas Habib Al Qutaifi⁵⁵. As the present use of constitutional adjudication illustrated, it is a reaction to such threats, it requires access to lawyers and the courts, and as Section V found, it gives benefits that are case-specific and not uniform among claimants in comparable situations. Each of these shortcomings could be addressed by codification, in the form of domestic law or binding administrative guidance, that would provide anticipatory rather than post hoc protection, would provide standardised protection instead of protection contingent on successful litigation, and, most importantly for the purposes of the argument advanced in this paper, allow protection under non-refoulement to be expressly and structurally linked to the CEDAW-informed persecution analysis developed in Section IV, rather than to the comparatively narrower language of Article 21, through which Indian constitutional jurisprudence has thus far evolved. In effect, codification would fulfil the interpretive project this paper has been advancing: it would embed it into a binding domestic instrument, rather than continuing to leave it hanging on the continued good faith of constitutional courts to import the non-refoulement obligations into Article 21⁵⁶.

These three changes cannot, however, be seen as spontaneous and self-fulfilling outgrowths of the reinterpretation of doctrine. This is the answer to the question that is stated in the abstract of the paper. Interpretative innovation, CEDAW’s constitutive incorporation into persecution analysis, provides the substantive substance of what persecution, well-founded fear and State protection mean in cases of sexual violence, but this does not mean it is always applied in the same way. Realising that content, in practice, also requires institutional infrastructure; trained adjudicators who are fluent in applying the due diligence framework developed under the CEDAW, as is currently the case with PSG; standardised procedures that do not vary in quality based on the resourcing of the specific UNHCR field office or domestic authority being engaged in the process; and, in jurisdictions such as India, a robust domestic legal framework in place that enables the operationalisation of obligations under the constitution and the treaty, currently only in a patchwork form judicially constructed through decision-making.

The doctrinal and institutional aspects of reform are best viewed as mutually constitutive, rather than sequential; doctrinal reform implies institutional reform, but the latter can also be the ground for reform of the first. Without institutional change, doctrinal reinterpretation may generate a more nuanced meaning of the law, with the same arguments now being applied in different ways, and thus different results to materially similar claims in different cases, by different caseworkers, offices, and jurisdictions. Efficiency, consistency and transparency in the application of an inadequate framework is not protection, and indeed risking the opposite outcome, institutional reform without doctrinal reorientation can succeed in producing more efficient, better resourced, and more procedurally consistent RSD systems, but in which the same categories in the Convention are applied without its substantive content. It is because each reform, as it were, is a part of a single whole that the two must be implemented simultaneously: institutional reform cannot be a second implementation detail, contingent on the success of doctrinal argument, but rather an indispensable and co-equal condition for the normative innovations advanced in this paper to be operative, rather than merely aspirational, for the survivors of sexual violence who seek protection under the 1951 Convention.

Conclusion

It has been claimed that the failure of the 1951 Refugee Convention to recognise sexual violence as persecution is not a judicial problem of lack of creativity, but rather a structural problem connected with the Convention’s gender-neutral formulation, the lack of consistency of “particular social group” as the main surrogate category for gender-based persecution claims and the under use of UNHCR’s own guidance and CEDAW’s General Recommendations. The Convention’s promise of protection is not yet fulfilled, not because the Convention’s language excludes survivors of sexual violence in South Asia, and the wider Global South from protection, but because the interpretive resources that could make the Convention’s promise of protection come to pass have not consistently or authoritatively been brought to bear upon it.

The main contribution made by the paper has been to suggest that CEDAW, and in particular the General Recommendations 32 and 35^57,58, should be interpreted as forming part of the substantive body of refugee law and not merely supplementary to the fundamental concepts of persecution, well-founded fear and State protection. This reorientation, based on the principle of systemic treaty integration under the Vienna Convention on the Law of Treaties⁵⁹, Article 31(3)(c), provides refugee law with a substantive and structural account of gender based violence that can’t be produced inhouse through the traditional gender-neutral approach to refugee law. When applied to Indian State practice, where there has been no specific domestic legislation governing complementary protection, but instead a patchwork of decisions on Article 21 and UNHCR mandate RSD, the reinterpretation shows the potential and current limitations of the current form of complementary protection.

The three reforms proposed, the Protocol on Gender Based Persecution by the UNHCR, the integration of CEDAW into the RSD procedure, and codified implementation of the non-refoulement protection, are not alternatives to an interpretation of CEDAW, but are essential institutional complement. This paper has tried to show that interpretative innovation and institutional reform go hand in hand; that is, neither can be successful without the other if the protection gap identified in this paper is to be filled. The point of this argument is very real for those who have suffered sexual violence in the Global South, and in India, in particular, and it provides a doctrinal route for protection that does not require that a formal amendment be made to the Convention, or that India accede to one of the instruments that it has, for seventy-five years, refused to join. This paper’s argument can be seen as part of the larger body of work on promoting substantive gender equality in international refugee law, not as an add-on to be fitted into the existing genres of “perception” and “protection,” but as an integral component of the dynamic of these concepts.

*****

Footnotes

1. Convention Relating to the Status of Refugees, Adopted by the United Nations Conference of Plenipotentiaries on 28 July 1951, entered into force 22 April 1954, 189 UNTS 137, Art. 1A(2).

2. UNHCR, Guidelines on International Protection No. 1: Gender-Related Persecution within the Context of Article 1A(2) of the 1951 Convention, 7 May 2002, UN Doc. HCR/GIP/02/01.

3. James C. Hathaway and Michelle Foster, The Law of Refugee Status, 2nd Edition (Cambridge University Press, Cambridge, 2014), pp. 185-207.

4. UNHCR, Guidelines on International Protection No. 2: “Membership of a Particular Social Group” within the Context of Article 1A(2) of the 1951 Convention, 7 May 2002, UN Doc. HCR/GIP/02/02.

5. Committee on the Elimination of Discrimination against Women, General Recommendation No. 32 on the Gender-Related Dimensions of Refugee Status, Asylum, Nationality and Statelessness of Women, 14 November 2014, UN Doc. CEDAW/C/GC/32.

6. Committee on the Elimination of Discrimination against Women, General Recommendation No. 35 on Gender-Based Violence against Women, Updating General Recommendation No. 19, 26 July 2017, UN Doc. CEDAW/C/GC/35.

7. Vienna Convention on the Law of Treaties, Adopted 23 May 1969, entered into force 27 January 1980, 1155 UNTS 331, Art. 31(3)(c).

8. 1951 Convention, note 1.

9. Hathaway and Foster, note 3.

10. Protocol Additional to the Geneva Conventions of 12 August 1949, and Relating to the Protection of Victims of International Armed Conflicts (Protocol I), Adopted 8 June 1977, entered into force 7 December 1978, 1125 UNTS 3, Art. 76(1); Protocol Additional to the Geneva Conventions of 12 August 1949, and Relating to the Protection of Victims of Non-International Armed Conflicts (Protocol II), Adopted 8 June 1977, entered into force 7 December 1978, 1125 UNTS 609, Art. 4(2)(e).

11. Prosecutor v. Akayesu, Case No. ICTR-96-4-T, Judgment, International Criminal Tribunal for Rwanda, 2 September 1998, paras. 596-598, 731-734.

12. Prosecutor v. Kunarac, Kovač and Vuković, Case Nos. IT-96-23-T & IT-96-23/1-T, Judgment, International Criminal Tribunal for the former Yugoslavia, 22 February 2001, paras. 436-460; affirmed in Prosecutor v. Kunarac, Case Nos. IT-96-23 & IT-96-23/1-A, Appeal Judgment, 12 June 2002.

13. Rome Statute of the International Criminal Court, Adopted 17 July 1998, entered into force 1 July 2002, 2187 UNTS 3, Arts. 7(1)(g), 8(2)(b)(xxii), 8(2)(e)(vi).

14. Committee against Torture, General Comment No. 2: Implementation of Article 2 by States Parties, 24 January 2008, UN Doc. CAT/C/GC/2, para. 18; Aydın v. Turkey, App. No. 23178/94, European Court of Human Rights, 25 September 1997.

15. CEDAW General Recommendation No. 32, note 5.

16. CEDAW General Recommendation No. 35, note 6.

17. Deborah E. Anker, “Refugee Law, Gender, and the Human Rights Paradigm”, Harvard Human Rights Journal, vol. 15 (2002), pp. 133-154 at p. 143.

18. Islam (A.P.) v. Secretary of State for the Home Department; R. v. Immigration Appeal Tribunal and Another, Ex Parte Shah (A.P.), [1999] UKHL 20, [1999] 2 AC 629.

19. Matter of Kasinga, 21 I&N Dec. 357 (BIA 1996).

20. Canada (Attorney General) v. Ward, [1993] 2 SCR 689 (Supreme Court of Canada).

21. 1951 Convention, note 1.

22. Ward, note 20.

23. Matter of Acosta, 19 I&N Dec. 211 (BIA 1985).

24. Applicant A v. Minister for Immigration and Ethnic Affairs, (1997) 190 CLR 225 (High Court of Australia).

25. UNHCR Guidelines No. 2, note 4.

26. Islam and Shah, note 18.

27. Heaven Crawley, Refugees and Gender: Law and Process (Jordan Publishing, Bristol, 2001), Chapter 3.

28. UNHCR, “Note on Burden and Standard of Proof in Refugee Claims”, 16 December 1998, para. 11.

29. Vienna Convention on the Law of Treaties, note 7.

30. UNHCR, Handbook on Procedures and Criteria for Determining Refugee Status and Guidelines on International Protection under the 1951 Convention and the 1967 Protocol relating to the Status of Refugees, UNHCR, Geneva, 2019 Reissue, UN Doc. HCR/1P/4/ENG/REV.4, paras. 195-205.

31. Convention on the Elimination of All Forms of Discrimination against Women, Adopted 18 December 1979, entered into force 3 September 1981, 1249 UNTS 13, Art. 1.

32. Committee on the Elimination of Discrimination against Women, General Recommendation No. 19: Violence against Women, Eleventh Session, 1992, UN Doc. A/47/38.

33. CEDAW General Recommendation No. 35, note 6.

34. CEDAW General Recommendation No. 32, note 5.

35. CEDAW General Recommendation No. 35, note 6.

36. UNHCR Handbook, note 30.

37. Crawley, note 27.

38. Jane Freedman, Gendering the International Asylum and Refugee Debate, 2nd Edition (Palgrave Macmillan, London, 2015), pp. 96-118.

39. The Foreigners Act, 1946 (India), s. 3.

40. The Registration of Foreigners Act, 1939 (India).

41. National Human Rights Commission v. State of Arunachal Pradesh, (1996) 1 SCC 742.

42. Louis De Raedt v. Union of India, (1991) 3 SCC 554.

43. Ktaer Abbas Habib Al Qutaifi v. Union of India, 1999 Cri LJ 919 (Gujarat High Court, 12 October 1998).

44. For details pertaining to UNHCR’s operations in India, see UNHCR, “Fact Sheet: India” (2025), available at: <https://www.unhcr.org/in/about-unhcr/where-we-work/asia-and-pacific/india> accessed on 10 July 2026.

45. The Constitution of India, 1950, Art. 21.

46. For details pertaining to India’s ratification of the Convention on the Elimination of All Forms of Discrimination against Women on 9 July 1993, see United Nations Treaty Collection, Multilateral Treaties Deposited with the Secretary-General, Ch. IV.8, available at <https://treaties.un.org/Pages/ViewDetails.aspx?src=TREATY&mtdsg_no=IV-8&chapter=4&clang=_en> accessed on 10 July 2026.

47. The Foreigners Act, 1946, note 39.

48. UNHCR Guidelines No. 1, note 2.

49. UNHCR Guidelines No. 2, note 4.

50. CEDAW General Recommendation No. 32, note 5.

51. CEDAW General Recommendation No. 35, note 6.

52. UNHCR Guidelines No. 2, note 4.

53. United Nations Treaty Collection, note 46.

54. NHRC v. State of Arunachal Pradesh, note 41.

55. Ktaer Abbas Habib Al Qutaifi, note 43.

56. Constitution of India, note 45.

57. CEDAW General Recommendation No. 32, note 5.

58. CEDAW General Recommendation No. 35, note 6.

59. Vienna Convention on the Law of Treaties, note 7.