Unprincipled and unrealised: CEDAW and discrimination experienced in the context of migration control

This article analyses the CEDAW Committee’s General Recommendations and Views on individual complaints, to evaluate its contribution to the elimination of discrimination against women experienced in the context of migration control. It makes two arguments. First, the Committee’s General Recommendations contain a range of doctrinal and empirical shortcomings. This opacity, and these omissions, considerably reduce the value of the Committee’s statements as a means by which States’ discriminatory migration control practices might be contested. Second, the Committee’s decisions, in communications concerned with discrimination experienced in the context of migration control, are inconsistent with those standards that it has set, and with the decisions it makes in other types of cases. A detailed analysis of the jurisprudence grounds the conclusion that the Committee is, in practice, according States a margin of appreciation that varies according to the subject of the complaint. Particular, representative communications are drawn on to argue that the margin granted in cases concerned with migration control is over-wide, characteristic not of appropriate (quasi) judicial restraint, but unprincipled deference. The article concludes by suggesting how some of the criticisms outlined may be remedied, notably by the Committee adopting its own justification and proportionality assessment.


Introduction
The Convention on the Elimination of All Forms of Discrimination against Women (CEDAW) is one of nine core international human rights instruments. Adopted in 1979, it has, at the time of writing, 189 States parties. The Convention is monitored by the Committee on the Elimination of Discrimination against Women (the Committee). Committee members are 'experts of high moral standing and competence in the field covered by the Convention', who serve in their personal capacities. 1 The Committee reviews States' reports on the 'legislative, judicial, administrative or other measures' they have adopted to give effect to the Convention. 2 It also issues General Recommendations and, following the adoption of an Optional Protocol, hears communications from individuals who believe that their CEDAW-protected rights have been violated. 3 As will be apparent from its title, CEDAW's protection from discrimination on the grounds of sex/gender is asymmetrical. 4 By defining and prohibiting discrimination against women, guaranteeing women's political participation and reproductive autonomy, and securing women's rights in law, education, employment, cultural and family life, the Convention seeks to ensure women's 'exercise and enjoyment of human rights and fundamental freedoms…'. 5 That CEDAW does not protect all people from discrimination on the grounds of sex/gender, does not mean that it cannot, or should not, protect a significant proportion of them. 6 Principled and pragmatic reliance on a legal regime which ensures that women's rights are human rights does not preclude, or undermine, the adoption and use of a range of instruments to remedy other, and all forms of gendered inequality. 7 Feminist and legal reliance on the 'non-monolithic' category of 'woman' 8 remains a 'necessary, strategic and non-dangerous form of weak essentialism' 9 which 'serves to underscore that worldwide human beings experience privation at a disproportionate rate because they are women.' 10 The Convention, and key terms within it, may be interpreted purposively, contributing to the development of a more 'liberatory and inclusive conception of gender in international (and domestic) law.' 11 This, and the Convention's mobilisation in conjunction with discrimination prohibitions that encompass multiple, and intersecting grounds, ensure that the Convention remains a vital, but by no means the only, response to sexed/gendered discrimination. 12 In this article, I analyse the Committee's General Recommendations and Views on individual complaints, to evaluate its contribution to the elimination of discrimination against women in a particular legal context -migration control. In Part I, I review all the Committee's General Recommendations in order to identify doctrinal and empirical shortcomings and omissions in the Committee's reasoning on when, and why, certain widespread migration control practices are discriminatory. My starting point is the position, in international law, that while the Committee's interpretation of CEDAW is authoritative, 13 it is only correct when in accordance with the ordinary meaning of relevant terms in their context, and in light of the Convention's object and purpose. 14 In Part II, I examine the Committee's application of those obligations it has delineated to States, reviewing individual complaints brought under the Convention's Optional Protocol over a 9-year period. This analysis reveals two disparities, or inconsistencies. The first is between how the Committee approaches migrant women's rights in abstract, when discharging its general interpretative role, and how it behaves when confronted, in a communication, with a particular rule or practice that is, on its face, discriminatory. The second, and related disparity, is between how the Committee approaches communications that allege discrimination experienced in, and out, of the migration control context. I explain these disparities by arguing that the Committee accords a margin of appreciation to States. The Committee has not generally adopted the margin of appreciation doctrine, or set out the factors that determine how it is applied in different types of cases. 15 In practice, however, the Committee is indeed granting a margin of appreciation to States, one that is significantly wider in communications where migration controls are challenged, than in communications where they are not. ColmÓ Cinnéide has argued that using international human rights law to contest discrimination experienced in the context of migration law and control is challenging, and often futile. 16 My analysis substantiates and developsÓ Cinnéide's critique. By explaining that the Committee does in fact accord a margin of appreciation, and in concluding that it should not do so, this article also provides a richer understanding of how the Committee does, and should, operate as a soft court.
The arguments I advance here rest on three premises: first, that CEDAW and its Committee play a significant role in the international protection of human rights. Second, that discrimination along multiple and intersecting axes, including against women, is 'rife' 17 in the migration control context, and that third, CEDAW is uniquely placed to contest such disadvantageous treatment.
Both Theodor Meron and Loveday Hodson characterise CEDAW (with or without its Optional Protocol) as a 'second-class' instrument. 18 In contrast, I situate this article within the literature that highlights the transformative contribution that the Convention, and its Committee, has made to human rights law and protection. The Convention is a 'forceful statement' that discrimination against women is unacceptable, one that States in all regions of the world have agreed to. 19 It is a 'catalyst for change', which has been relied upon 'in multiple ways by women across the world' in accordance with what is 'best suited to their local contexts.' 20 The Committee sets legal standards. In its General Recommendation 19 (GR 19), the Committee's definition of gender-based violence as discrimination and a violation of women's human rights 21 closed a normative gap in international law, 22 in addition to informing the approach taken to such violence by the broader UN. 23 The Convention and Committee have inspired and rooted regional efforts to respond to violence and discrimination against women, a 'growing synergy' being identified between them and the Council of Europe in this regard. 24 The Convention has also been widely cited by national courts, 25 Fareda Banda's 'snapshot' of CEDAW in the African context detailing, for example, its use in judgements, as well as in legal education and advocacy. 26 Finally, by engaging with States and NGOs during the reporting process, the Committee also has a 'culturally constitutive role', one that goes beyond interpretation and adjudication to create meaning on gender and equality. 27 My second premise is that women are particularly disadvantaged by migration law and control. Such disadvantage may be the product of rules or practices which explicitly treat women less favourably than men, by, for example, restricting or preventing emigration. 28 Usually, however, it flows from rules which are, on their face, gender neutral, but which actually disadvantage women, or particular groups of women. Such disadvantage is expressive and material, entailing stereotyping, the devaluation of women's work, the loss of employment or other opportunities, and the confinement of women to certain roles and relationships. 29 It may involve an increase in women's extant risk of experiencing violence, or it may also simply be violent. 30 Third, CEDAW is, theoretically at least, institutionally and doctrinally well placed to remedy such disadvantageous treatment. CEDAW's article 1 definition of 'discrimination against women' encompasses both direct and indirect discrimination. 31 Indeed, adverse effect, whether material or expressive, intentional or otherwise, is the definition's 'critical criterion'. 32 It may, therefore, be relied on to challenge rules which explicitly disadvantage women on the grounds of sex/gender, and those which negatively affect women, even though they appear to be gender-neutral. The Committee's characterisation of gender-based violence as unlawful discrimination brings any migration control practice that entails or involves it, including push-backs and detention, within its purview. It also enables the Committee to scrutinise rules that increase women's vulnerability to violence perpetrated by non-State actors, such as those that produce destitution. CEDAW's definition of discrimination against women reaches out of the Convention, to apply to rights protected in other human rights instruments, and under customary international law. 33 This enables article 1 to reinforce and complement two regimes relevant to migration whose key instruments fail to explicitly prohibit sex/gender discrimination: the international protection and trafficking regimes. 34 It also requires States to provide gender-sensitive interpretations of those instruments. Unlike other conventions which, in different ways and to different extents, differentiate between the rights of citizens and noncitizens, 35 the CEDAW Committee is clear that the obligations of States parties apply: without discrimination both to citizens and non-citizens, including refugees, asylum-seekers, migrant workers and stateless persons, within their territory or effective control, even if not situated within the territory. States parties are responsible for all their actions affecting human rights, regardless of whether the affected persons are in their territory. 36 The discrimination women face 'based on sex and gender is inextricably linked with other factors that affect women, such as race, ethnicity, religion or belief, health, status, age, class, caste and sexual orientation and gender identity'. 37 Recognising this, the Committee has affirmed that intersectionality is a 'basic concept' for understanding States' obligations under the Convention. 38 Within the context of migration control, particular instantiations of these links may include the co-constitution of discrimination on the grounds of nationality, race and sex/gender. Crossing a border in pursuit of protection from, for example, persecution on the grounds of gender identity, may lead to a lack of recognition for, or even the erasure of, that identity. 39 The Committee's intersectional approach, including its insistence that refugees and migrants be identified as rights-bearers, 40 is, therefore, essential to reveal and remedy the harms faced by particularly disadvantaged and persecuted groups of women. Finally, the Committee has, as the above quote illustrates, articulated a broad, effects-based approach to jurisdiction under the Convention. 41 CEDAW may, therefore, be relied on to challenge the discriminatory treatment of non-citizen women in a variety of different contexts, including those where the harm feared, or women concerned, are out of the relevant State's territory.
Delineating States' obligations: The CEDAW Committee's interpretation of the Convention Having defined discrimination against women in article 1, the Convention requires States to prohibit it (article 2(b)); to establish women's equal legal protection (2(c)); to refrain from discriminating against women (2(d)); to eliminate discrimination from other sources (2(e)); and to modify discriminatory laws or practices (2(f)). The Committee has drawn on articles 1 and 2, in conjunction with other CEDAW provisions, 42 to outline, in its General Recommendations, a series of mostly mutually reinforcing obligations which seek to identify, prohibit and establish remedies for unlawful discrimination in States' migration control practices. Relevant obligations can be identified in almost all of the Committee's Recommendations, including those concerned with women's health (GR 24), the rights of migrant workers (GR 26), the rights of women in conflict (GR 30), the gender-related dimensions of refugee protection and statelessness (GR 32), violence against women (GR 19, strengthened by GR 35) and trafficking (GR 38). 43 Pursuant to these, States should, for example, reform indirectly discriminatory rules, including those that impose minimum income requirements, which require language proficiency or economic self-sufficiency. 44 States should also revise rules that disproportionately disadvantage women, like those that make spouses, partners and particular types of labour migrants dependent on another. 45 They should also provide gender-sensitive refugee recognition and trafficking identification procedures. Such procedures should themselves draw on, and apply, gendersensitive interpretations of relevant international obligations. 46 In these recommendations, the Committee has also, however, made a range of interpretative statements that either do not explain why rules it characterises as unlawfully discriminatory are so, or which provides reasoning on this point that is partial, or unclear. It has, for example, stated that family reunification rules for migrant workers should not directly or indirectly discriminate against women. 47 Earlier in the relevant recommendation, the Committee expressed its concern about migration statuses that are particularly disadvantageous because they restrict their holder's ability to be accompanied by family members. These statuses, the Committee opines, are often applied to particular types of 'female-dominated' employment, such as domestic work. 48 What the Committee omits is any explanation that would enable a State, or the potential author of a communication, to assess whether such apparently gender-neutral restrictions disadvantage women in any particular context or jurisdiction, whether directly or, as is most likely to be the case, indirectly. Indeed, the Committee does not provide any legal argumentation which connects its particular concerns about restrictions, which produce gendered disadvantage, and its general assertion that family reunion rules should not discriminate.
This normative under-development has two interrelated elements. 49 The first, as the above example illustrates, relates to the content of the relevant non-discrimination provision. It concerns the Committee's failure to delineate precisely what obligations States assume in relation to their family reunion, or other rules. Can differential treatment be justified? And if so, how? What constitutes a legitimate aim? How should the relationship between that aim, and the means employed to secure it, be assessed in the migration control context? 50 The second element relates to how the breach of any particular obligation may be established. Can statistical information, for example, be relied on to prove differential treatment? 51 The Committee does not provide answers to these, or to related, evidential questions. This criticism of the Committee, does, of course, presuppose that its function is to establish such normsthat the Committee should, for example, articulate a clear position on justification. The Committee is not an international court or tribunal, so this line of argumentation itself requires some justification. Those bodies that are, such as the ICJ, 'ascribe great weight' to the recommendations of United Nations Treaty Bodies (UNTBs). 52 They do so first, because they are the independent bodies, established to supervise 'their' Conventions, and second, because it is necessary to: achieve the necessary clarity and the essential consistency of international law, as well as legal security, to which both the individuals with guaranteed rights and the States obliged to comply with treaty obligations are entitled. 53 The Committee's failure to provide those reading its recommendations with the legal tools necessary to identify and contest discrimination adversely affects both individuals who fear their Convention-protected rights have been violated, and the broader international rule of law.
The CEDAW Committee's failure to articulate norms in the migration control context is one that it shares with the Human Rights Committee (HRC) and the Committee on the Elimination of Racial Discrimination. 54 It has also, however, engaged in a distinctive kind of normative overreach. A chronological reading of its recommendations reveals a shift in the Committee's interpretative approach and focus, from one that identifies obligations as arising from specific Convention provisions, to one which does not always, or explicitly, make this link, and which purports to provide a 'gender-sensitive interpretation' of international law. 55 The criticism being levelled at the Committee here is not with its adoption, over time, of a more expansive role and remit. Indeed, its recommendations on gender-sensitive interpretations of the refugee definition, and gender-sensitive determination procedures, demonstrate the normative and practical value of bringing a substantive conception of non-discrimination into dialogue with the law on international protection. Rather, it is that the Committee's failure to explain what constitutes discrimination is even more apparent in these later, un-moored recommendations. The Committee has, for example, stated that pushbacks and detention increase women's vulnerability to exploitation, a conclusion that is consistent with a wealth of empirical research. 56 When, however, are such practices, which are so widely used by States, unlawfully discriminatory? Always? Or only when they are implemented as part of 'discriminatory migration and asylum policies' (itself unexplained)? 57 Or, when States have failed to implement the corresponding recommendation to provide 'increased access to pathways for safe and regular migration'? 58 What is sufficient in this regard, just an increase in (any) pathway, or that women and men have equal access to a range of specific migration opportunities? And how could, and should, this be assessed? Several of the Committee's statements on trafficking, and States' article 6 obligations to suppress it, lack precision, or indeed, any clear legal foundation. Recommendations, such as 'promoting peace', responding to climate change, and enabling 'lifelong learning', 59 are simply not consonant with the ordinary meaning of article 6, both in its relevant context, and in light of CEDAW's object and purpose. 60 If this is incorrect, and these recommendations are legally sound interpretations of the Convention, it is unclear what States need to do to implement them. While, therefore, the Committee has identified and outlined a range of pertinent obligations, its unique combination of normative under-development, and overreach, has significantly hampered its ability to eliminate discrimination in migration control.

Holding States responsible: The Committee's Views on communications brought under CEDAW's Optional Protocol
To understand how the Committee applies those obligations it has identified to States, I compiled a database of 9 years of communications concerning discrimination and genderbased violence, experienced both within and outside of the migration control context. 61 The Committee's adjudication of communications has, broadly speaking, two stages, with first, admissibility, and second, merits being determined. Communications are admissible where the author has exhausted domestic remedies, and sufficiently substantiated their complaint. 62 The standard of review that the CEDAW Committee applies is a modified version of the HRC's, the Committee, approaching complaints on the basis that it is generally for the authorities of States parties to the Convention to evaluate the facts and evidence or the application of national law in a particular case, unless it can be established that the evaluation was biased or based on gender stereotypes that constitute discrimination against women, was clearly arbitrary or amounted to a denial of justice. 63 Comparing the Committee's decisions in cases outside, and in, the migration control context, reveals a significant difference between how the Committee discharges its adjudicative role in different types of cases. To foreshadow the more detailed discussion of the Committee's reasoning that follows, in the first group of cases, its examination of allegations of discrimination outside of the migration control context is relatively careful, States' conduct being subject to detailed scrutiny by reference to its General Recommendations. In the second, concerning allegations of discrimination within the migration control context, scrutiny is much more cursory, the Committee accepting, for example, States' determinations of women's asylum claims without examining whether or not their procedures comply with the standards set in GR 32. Indeed, the Committee defers to States' assessments even where the evidence before them indicates that the relevant process does not meet those standards. This approach insulates States from challenge, denying women a forum to contest breaches of the Convention. This assessment is more critical than that reached by others who have analysed the Committee's decisions. Loveday Hodson has, for example, simply stated that the Committee has 'been unable to establish much of a voice on the asylum claims of vulnerable women.' 64 My conclusion is based on an analysis of a series of decisions made over a particular, defined, period of time. It may, therefore, be unrepresentative of the Committee's work over the longer term. The communications that concerned migration control were also from a much narrower selection of jurisdictions than those which did not, a geographical skew that may have produced substantive distortions in the communications I reviewed. 65 The inconsistency with which the Committee interprets, and then applies, rights to non-discrimination and equality in different legal contexts has, however, been identified by others. Simone Cusack and Lisa Pusey explain the sometimes-wider, sometimes-narrower gap between obligation and its application, as being attributable to the Committee's failure, in certain types of cases, to take a sufficiently robust gender analysis. 66

I think that Başak Çalı, Cathryn Costello and Stewart
Cunningham are closer to the mark in their conclusion that it has taken a 'tentative, deferential approach in its non-refoulement cases.' 67 I contend that this deference can be explained by reference to the Committee's granting of an overly-wide margin of appreciation to States in communications concerned with migration control.
The margin of appreciation is a practice, or doctrine, of judicial self-restraint that involves assigning weight in international adjudication to a respondent State's reasoning on the basis of certain factors. 68 A judicial creation of the European Court of Human Rights (ECtHR), the doctrine has since been endorsed by States in the region. 69 It has also been adopted by other international courts and tribunals, including the Inter-American Court of Human Rights. 70 The granting of a margin of appreciation to States has been defended on a number of grounds. It demonstrates respect for the democratic legitimacy of certain types of decisions, gives effect to the principle of subsidiarity, ensures institutional competence and inter-institutional clarity, and enables the establishment of minimum human rights standards. 71 The standard of review, and the granting of a margin of appreciation to States, go hand-in-hand. The higher the standard, the narrower the margin of appreciation. The margin granted will depend on factors including the quality of parliamentary and judicial oversight, 72 and the discrimination ground at issue. 73 Where the rights of a vulnerable group are at issue, one that has been subject to entrenched discrimination, the margin of appreciation accorded to States should be narrow. 74 Scholarship on the ECtHR's application of the margin of appreciation has highlighted its unprincipled, or overly-wide application, particularly in the context of transnational migrants' and religious minorities' rights. 75 In addition to justifying the doctrine's use by the ECtHR, Dominic McGoldrick has advocated for its formal adoption by the HRC. 76 The HRC, while not using the language of the doctrine, appears to rely on it in practice. 77 Space precludes me from bringing criticisms of the doctrine's use by the ECtHR, and McGoldrick's arguments on the HRC, into conversation with each other here. I refer to them to situate my argument that if the Committee is, like its UNTB cousin the HRC, according a margin of appreciation in practice, it should not do so. Alternatively, if the Committee does wish to use the doctrine, it should adopt it explicitly and ensure its principled, properly calibrated, application.
The 9 years of communications analysed concerned discrimination and gender-based violence experienced outside of, and within, the migration control context. The first group of cases did not raise any migration or related issues. It comprised 36 communications concerning discrimination and gender-based violence. The Committee found the majority of these cases admissible, identifying breaches of the Convention in all those communications whose merits it considered. 78 It also subjected respondent States' conduct to a high level of scrutiny, applying the Convention, and relevant recommendations on violence against women, robustly. In ON and DP v the Russian Federation, 79 for example, the Committee recognised and responded to the intersectional violence that lesbian women may be subject to, identifying failures and making recommendations at both the individual and structural level. 80 The Committee has found States responsible for failures to investigate and prevent violence against women, 81 emphasising the obligations owed to particularly vulnerable women, 82 and to children. 83 Overall, and as Andrew Byrnes and Eleanor Bath have also concluded, in communication after communication, the Committee required States not just to provide an appropriate legal framework to combat gender-based violence, but also to afford such protection in practice. 84 During the same period, the Committee heard 34 communications concerned with discrimination and violence experienced in the context of migration control. In contrast to the first group, the Committee found the majority of these communications inadmissible. 85 When doing so, it subjected States' conduct to a much lower level of scrutiny, as a brief discussion of two pertinent cases demonstrates.
The author of MEN v Denmark fled Burundi, describing both the political persecution she feared as an opposition party member and the rapes she had been subject to when attempting to escape it. 86 MEN's communication was found inadmissible by the majority of the Committee on the grounds that the author had not exhausted domestic remedies. They justified their decision by asserting that as MEN 'believed that it was coincidental that she had been raped…. the State party's authorities clearly had no opportunity to consider her gender-based violence allegations.' 87 The (in)admissibility of MEN's communication turned, therefore, on whether describing having been raped by three men at knife-point, but not defining that treatment as sex discrimination, was sufficient to raise such discrimination in her asylum claim. The majority of the Committee found that it was not. 88 In a powerful dissenting opinion, Ms DubravkaŠimonović found MEN's communication admissible, and identified a number of Convention violations. 89Š imonović, who went on to be appointed UN Special Rapporteur on violence against women, rejected the Committee's requirement that MEN explicitly characterise her treatment as sex discrimination. Instead, and drawing on GR 19,Šimonović explained that rape is 'universally accepted' to be a form of discriminatory, gender-based violence. 90 The failure of the majority to draw on its own recommendations when reviewing the State's determination of MEN's asylum claim is striking when compared to bothŠimonović's dissent, and the Committee's approach to sexual violence in non-migration cases.
MEN's communication was decided before the Committee adopted GR 32 (in 2014), on the gender-related dimensions of refugee status, or GR 35 (which updated and strengthened GR 19, in 2017) on gender-based violence. KIA v Demark, decided in 2019, concerned a Palestinian refugee and Jordanian national. KIA said that she had been 'tricked' into leaving Denmark, where she was resident as a teenager, and forced into marriage in Jordan where, for 11 years, she was subject to domestic and sexual violence. 91 Her claim for asylum was refused because she was inconsistent and, consequently, held to lack credibility. 92 KIA's communication was also found inadmissible, the Committee accepting the State's position that she had provided no evidence of procedural defect or arbitrariness. 93 The Committee referred to GR 32 just twice in the decision. It did not engage with pertinent recommendations on trauma or credibility assessment, or apply them to KIA's case. Other examples of communications where the Committee did not scrutinise States' asylum determination procedures by reference to the relevant recommendations include LO et al. v Switzerland (which concerned non-refoulement and domestic violence), 94 FHA v Denmark (non-refoulement, domestic violence and forced marriage), 95 and ARI v Denmark (stigma following sexual violence, risk of so-called 'honour-killing). 96 Of the nine communications in the migration control group whose merits were considered, the Committee identified breaches of the Convention in just three. Two of these concerned non-refoulement, 97 the third the rights of a migrant woman and her child who were subject to domestic violence. 98 Once again, the Committee's determination of these cases differs sharply from those that did not challenge migration controls. A brief consideration of one communication, where no Convention breach was found, serves to illustrate this.
The author in Rahma Abdi-Osman v Switzerland, 99 was kidnapped by a member of Al-Shabaab and forced to marry her abductor. She was held captive and raped on numerous occasions, becoming pregnant three times. A child born as a result of one of these pregnancies was taken from her. Her further two pregnancies were ended by force. 100 Ms Abdi-Osman fled Somalia and eventually claimed asylum in Italy. She was granted subsidiary protection following a process she alleged was flawed, and during which she was homeless and experienced further sexual violence. Ms Abdi-Osman then married an Italian national of Somali origin who was temporarily admitted to Switzerland. She joined her husband there and claimed asylum, a claim that was refused. Italy agreed to readmit Ms Abdi-Osman, the Swiss authorities driving her to the border and leaving her there. Without information or financial support, Ms Abdi-Osman was homeless again for 12 days, after which she returned to Switzerland. 101 There, following a further deterioration in her health, she made another claim for asylum, which was again rejected. At this point, Ms Abdi-Osman, who was pregnant, faced being returned to Italy again, this time with a new-born baby. She brought a communication to the Committee, alleging that doing so would expose her to further violence, and to significant harm, as her complex health needs would not be met. The Committee, however, found nothing that: amounted to, or provoked, discrimination or rendered decisions made by authorities arbitrary in the author's case. Moreover, it is for each sovereign State party to determine the nature, structure and procedures of its own asylum system, as long as basic procedural guarantees set down in international law are provided. 102 In GRs 19, 32 and 35, the Committee has set the international standards against which States' responses to the violence asylum-seeking women have experienced, or are at risk of, should be evaluated. While the Committee refers to these recommendations, it does not, in its Views, draw on their substance when reviewing the relevant States' conduct. For example, the Italian authorities do not appear to have identified Ms Abdi-Osman as a victim of torture and sexual violence, or provided her with safe accommodation, health services and adequate support, as required by GR 32. 103 Rather than evaluate Italy's treatment of Ms Abdi-Osman, and Switzerland's decision to return her, by reference to these recommendations, the Committee 'notes' Switzerland's assertion that there are 'no specific indications that the author was not afforded protection in Italy in the past'. 104 This is despite the fact that the lack of protection and health care that Ms Abdi-Osman feared return to is precisely what occurred on not one, but two occasions. We also have no idea whether, during the asylum process, Ms Abdi-Osman was provided with competent legal representation, a trauma-sensitive interview, and psychosocial counselling. 105 This is the case even though GR 32 describes such protections as 'rights', rooted in States' CEDAW article 2(c) obligations. 106 Finally, the Committee makes no effort to determine whether or not the relevant protections were in place to enable the sexual violence Ms Abdi-Osman experienced in Italy to be reported, and then investigated. 107 The Committee's decision left Ms Abdi-Osman, a woman who had significant health problems as a result of torture and sexual violence, to be deported with a baby to a country which had repeatedly failed to meet the standards the Committee has itself set. The Committee's failure to scrutinise her treatment by reference to its own recommendations, is not anomalous, but representative of others decided in the same period. 108 Comparing the preceding section's analysis, with this review of the jurisprudence, reveals an inconsistency between the standards the Committee has set on discrimination in migration control, and its application of those standards in communications which disclose it. This, and the deference the Committee shows to States when deciding migration communications, can be understood by reference to its granting, in practice, of a margin of appreciation that varies according to the subject of the complaint. In the group of communications that raise discrimination and violence, but which do not challenge States' migration controls, a high standard of review is applied, and the margin granted is correspondingly narrow. Breaches of Convention rights are identified, and conclusions about protection and reparation made. In comparison, and conversely, where cases challenge discrimination experienced in the context of migration control, the Committee's standard of review is much lower, and a significant margin of appreciation is accorded to States. Recommendations may be cited in passing, but they are rarely applied. Communications like those brought by MEN, KIA and Ms Abdi-Osman, indicate not only that a margin is applied, but that it is over-wide, characteristic not of appropriate (quasi) judicial restraint, but unprincipled deference.

Conclusion
The CEDAW Committee is indeed well-placed to respond to migration control's disadvantageous treatment of women. The Convention's definition of discrimination is capacious, covering direct and indirect discrimination, including that which is violent, whether experienced in the public or private domains. The Committee's broad, effectsbased approach to jurisdiction enables it to hold States responsible for their extraterritorial, discriminatory conduct. The Committee has, to an extent, capitalised on this advantageous positioning, impugning as discriminatory a myriad of rules and practices widely used by States. It has, for example, required States to reform indirectly discriminatory rules, abandon discriminatory practices, and provide gender-sensitive determination procedures that apply gender-sensitive interpretations of international protection and trafficking obligations.
There are, however, significant flaws in some of the reasoning the Committee has employed when delineating States' obligations. These include failing to explain how certain forms of discrimination should be established, doctrinally and empirically. Questions of justification are of fundamental significance in many types of discrimination case. They are, however, particularly salient to challenges of differential treatment experienced in the context of migration control, an area where States' assertions of the legitimacy and necessity of the measures they employ generally prevail. As the Committee has increasingly 'unmoored' its recommendations from the Convention's provisions, such omissions, and opacity, have become more pronounced. The Committee's adoption of expansive and unreasoned recommendations has considerably reduced the value of its interpretative statements as a means to contest discrimination in border control. 109 Such deficiencies leave women who have been disadvantaged by migration laws and practices without the legal tools necessary to challenge them.
International instruments rarely set out the standard of review to be applied by their monitoring bodies to the conduct (or lack thereof) of respondent States. Courts and treaty bodies' adoption and use of such standards is, consequently, an expression of those entities' ability to establish their own procedures. 110 The same can be said of the margin of appreciation doctrine, its creation by one court, in one region of the world, and its adoption in others. There may be sound reasons for granting States a margin of appreciation in some regional, or international, adjudicative contexts. As discussed briefly above, there is an engaging body of scholarship concerned with, and advocating for, the doctrine's explicit adoption and use by a range of courts and monitoring bodies. 111 Some of the relevant arguments can either be met in other ways, however, or simply do not apply to UNTBs. Concerns about subsidiarity, for example, may be resolved with admissibility criteria that ensure that UNTBs, amongst others, only adjudicate on those cases that cannot be resolved by national courts and tribunals. States parties to CEDAW, and other international human rights treaties, are not all democratic, so a doctrine that recognises the legitimacy of some decisions taken in States that are, may lack applicability. Whether or not arguments in favour of UNTBs' explicit adoption of the margin are accepted, the Committee's current approach denies women protection, including from refoulement, from the mechanism established to uphold their rights.
What follows are some briefly sketched concluding thoughts on how some of the criticisms I have outlined could be remedied. First, and most obviously, the Committee could simply enforce its own standards. To rephrase Rahma Abdi-Osman, while States may establish their own asylum systems, they should do so in accordance with those protections, identified by the Committee in its recommendations, which guarantee women's equal access to them. Individual and structural failures to follow such recommendations should be dealt with robustly. Second, the Committee could adopt its own justification and proportionality analysis, to determine whether or not rules which produce particular disadvantage are unlawfully discriminatory. Moving in this direction would require the Committee to deal with some of the normative under-development discussed above. The Committee would have to, and as I have argued, should in any event, set out how differential treatment may be established and justified. Once it is, and whatever the legal context, the Committee could consider whether a rule complained of has a legitimate aim, and whether there is a 'reasonable relationship of proportionality' between that aim and the means employed to secure it. 112 Alternatively, or in addition to the above, if the Committee does wish to accord States a margin of appreciation, it should adopt the doctrine explicitly, and articulate the factors which determine its application. Relevant here is migrant women's vulnerability, and the entrenched discrimination they are subject to. States' migration controls are often not debated or scrutinised by democratic institutions, and they may also be insulated from judicial challenge. These, and other factors, would militate against a wide margin of appreciation being granted in migration and discrimination cases. CEDAW's potential to contest discriminatory borders may yet be realised. To do so, the Committee should delineate obligations with clarity and precision, and develop a principled approach to adjudication. The Committee would then be in a position to identify, and hold States responsible for, all forms of discrimination against women.