Protecting Refugees Inside, Protecting Borders Abroad? Gender in the EU’s Responses to the ‘Refugee Crisis’

Migration tends to be denoted as a crisis which needs a solution. The European Union has developed policies for dealing with this crisis internally, within its borders, and externally. Both the experiences of migrants and European Union policy responses are gendered and have gendered effects. This article analyses how the European Union refers to gender in its definitions of and responses to the crisis. Grounded in feminist policy analysis, I scrutinize European Union internal and external policies under its Agenda of Migration. The analysis finds that European Union internal crisis responses demonstrate a more comprehensive understanding of gendered vulnerabilities and a commitment to human right provisions. External crisis responses reduce gender considerations to refugee women and the policy objective of reducing refugee arrivals which leads to further curtailing refugees’ access to protection. Showing how the crisis transforms the very meaning and scope of gender considerations to various degrees, the article furthers insights on how the European Union’s normative commitments develop in times of crisis.


Building Bridges: Gender, Refugee Policies and the EU's Promotion of Gender Norms
The analysis presented here brings together and contributes to work on gender in forced migration, EU refugee policy -especially in times of crisis, and the EU's promotion of gender equality. Bringing these distinct bodies of literature into dialogue with one another enables us to better understand how gendered vulnerabilities are (re)shaped and addressed by EU refugee policies in times of crisis.
The insight that gender matters throughout all phases of the migratory journey is not a novel idea. Scholarly work and feminist norm advocates have highlighted the need to address gendered inequalities in refugee and asylum policies (Bosworth et al., 2018;Hyndman and Giles, 2011;Martin, 2010;Spijkerboer, 2000). Earlier work was largely concerned with the situation of refugee women and their supposed vulnerabilities (Freedman, 2008(Freedman, , 2010. Yet, recent studies now highlight how gender in interaction with other social categories like age, sexual orientation, social class, religion and the body shape inequalities within refugee populations (Krystalli et al., 2018;Vervliet et al., 2014).
However, the recognition that refugee policies can create, reinforce or diminish gendered inequalities did hitherto not play a major role in the studies of EU refugee policies. Scholars have mainly scrutinized EU external refugee policies with regard to their effects beyond EU borders (Lavenex and Schimmelfennig, 2009;Lavenex and Uçarer, 2004;Slominski, 2013), inter alia how such policies control access and deter irregular migration. However, how these policies mobilize gender and other social categories to differentiate within the group of 'refugees' or 'migrants' has only been of marginal interest (Gerard and Pickering, 2014).
Various scholarly works have demonstrated how migration towards the EU has been constructed as a security issue (Bigo, 2002;Guild, 2009;Huysmans, 2000;Léonard, 2009Léonard, , 2010. At the core of the migration-security nexus is the juxtaposition of the freedom of movement within the EU's Schengen space and a supposedly threatening EU-outside. Haddad (2008: 201) argues that this dichotomy drives the EU's efforts to externalize border control to regions of origin. According to her, 'a securitarian ethic is promulgated within the EU, while a protection ethic is spread outside the EU in countries or regions of origin. Or, to take this further, protection is exported in order to maintain security inside'. Although feminist research has shown how constructions of protection and security are highly gendered (Muehlenhoff, 2017;Shepherd, 2007Shepherd, , 2009Tickner, 1992), feminist analyses of the EU's refugee and asylum policies remain scarce. Stachowitsch and Sachseder (2019) demonstrate the potential of such analytical approaches: analysing how FRONTEX mobilizes gendered and racialized frames, they show how the EU agency constructs and legitimizes itself as a protector of Europe against the migrant 'other' (see also Hoijtink and Muehlenhoff, this issue). In a similar vein and in line with the rationale of this Special Issue, this article seeks to understand how gender considerations differ in EU internal and external policies.
Moments of perceived crisis and emergency offer a fruitful context for that, because they reinforce vulnerabilities and create an imperative to act (Panebianco and Fontana, 2018). Times of crisis also reconfigure and lay bare the discursive, policy and legal categorisations, central for the governance of migration (Allen et al., 2018;Pallister-Wilkins, 2018). Implicitly or explicitly, gender and other social categories play into such categorical distinctions that draw the boundaries between wanted and unwanted migration. As Allen et al. (2018: 219) state, 'the law lends consequence to elements of social identity -including race, gender and religion -by investing them with legal consequences of inclusion and privilege or exclusion and subordination'. Thus, investigating how and in which policy areas the EU mobilizes gender shows how gender and the EU's governance of inclusion and exclusion relate.
The present analysis also seeks to build on and contribute to existent scholarly work on the EU's actions in the so-called 'refugee crisis' or the 'CEAS crisis' (Niemann and Zaun, 2017). Trauner (2016), for instance, has shown how the economic and financial crises impacted EU asylum policies and led to a discrepancy between EU laws and Member States' actual practices. In their Special Issue on 'EU Refugee Policies and Politics in Times of Crisis', Niemann and Zaun (2017) included diverse approaches and empirical studies to explain the absence of effective cooperation in response to the increase in refugee arrivals to the EU (e.g. Ripoll Servent, 2018;Slominski and Trauner, 2018). Scholarly work has scrutinized the crisis' impact on the EU's normative commitments mainly with regard to its humanitarian responsibilities (Jeandesboz and Pallister-Wilkins, 2016;Moreno-Lax, 2018), yet without paying attention to how they intersect with the EU's self-understanding as a 'gender champion' (Woodward and Van der Vleuten, 2014: 68).
Assessing the EU's realization of such normative engagements has a longer tradition in other policy fields. Feminist policy studies have examined gender equality norms in EU internal policies (Lombardo et al., 2009;Walby, 2004) as well as different external policy fields such as enlargement (Bretherton, 2001), development (Debusscher and Van der Vleuten, 2012;Elgström, 2000) and security policy (Ansorg and Haastrup, 2018;Guerrina et al., 2018). The field of EU refugee policy has received little attention (for an exception, see Allwood, 2015), despite the increasing political relevance of this field.

Expanding Gender Considerations in Times of Crisis?
The analysis of whether and how the EU expanded gender considerations in a moment of perceived crisis demands reflection on (1) the relation of crisis and policy change, (2) the meaning of expanding or mainstreaming gender considerations and (3) the understanding of gendered vulnerabilities.
First, the 'EU Refugee Crisis' marks a critical juncture for reframing the EU's role, policy objectives and responsibilities in the field of refugee policies (Niemann and Zaun, 2017). While here I am mostly interested in the 'narration of crisis' (Hay, 1999: 322)how crisis is constructed in policy documents -it is imperative to note that crisis is real as an experience and in its consequences. As Hay (1999: 323) asserts, crisis is 'a lived experience, it is a politically mediated moment of decisive intervention and structural transformation'. According to Boin et al. (2009: 81), crises produce 'framing contests to interpret events, their causes, and the responsibilities and lessons involved in ways that suit their political purposes and visions of future policy directions'. Therefore, the following analysis of the Agenda on Migration pays attention to how the EU frames the causes of and solutions to the crisis, internally and externally.
Second, feminist scholarship argues that a simple reference to gender in a policy document does not mean that this policy comprehensively incorporated, or mainstreamed, gender in a transformative way (Caglar, 2013). Gender mainstreaming, according to the definition, implies 'the (re)organization, improvement, development and evaluation of policy processes so that a gender equality perspective is incorporated in all policies at all levels and at all stages, by the actors normally involved in policy making' (Council of Europe, 2019). In addition, as Lombardo and Meier (2006: 153) note, a feminist reading of gender mainstreaming demands 'tackling the multiple interconnected causes that create an unequal relation between the sexes' and 'a focus on gender not only on women'. A crucial component for gender mainstreaming to be transformative is to rethink 'policy ends and means from a gender perspective and prioritiz[e] gender over competing objectives'. Thus, feminist policy analysis highlights what gender means, what actions gender considerations envisage and how they relate to other policy objectives.
Finally, in contrast to notions of vulnerability that presume a per se and fixed disadvantage of mainly refugee women, I depart from an intersectional understanding of gendered vulnerabilities. The concept of intersectionality was originally coined by Kimberlé Crenshaw (1991) who showed how different axes of inequality co-create a particular instance of marginalization. Intersectionality recognizes gendered vulnerabilities as contingent, relational and non-additive (Hancock, 2016). The way gender and other social categories create positions of relative privilege or marginalization is always in relation to the specific context and other subject-positions. For example, in a refugee camp, a young single mother may be particularly vulnerable, while in a European resettlement programme, single, heterosexual men may be underprivileged because they have little chance to be resettled (Turner, 2017). Depending on the context, gender and sexuality play out differently. Turner notes that 'a person is not vulnerable because they are a man or a woman, but because of what being a man or a woman means in particular situations' (Turner, 2016). Based on these conceptual reflections, the following section explains the details of the analytical approach.

Analysing Gender in the 'Refugee Crisis'
In this analysis, the EU Agenda on Migration together with all legislative documents is taken as an entry point to understand how problem representations and proposed solutions are constructed and gendered. As of June 2018, the Commission has issued 255 legislative documents, including annexes, summaries and staff working documents.
With a gendered perspective on problem representations and proposed policy responses, feminist adaptations of frame analysis (Lombardo et al., 2009) are well suited to analyse gender consideration in EU policies. Verloo's (2005Verloo's ( , 2007 Critical Frame Analysis (CFA) proposes to analyse the discursive framing of gender and its 'multiple meanings'. A set of synthesizing questions helps to analyse policy documents with regard to the problem definition (diagnosis: What is represented as a problem? Who is seen as responsible for the problem?), proposed solutions (prognosis: What should be done? Who is acted upon? How is action legitimized?) and normativity (What is seen as good and bad?). Such an analysis of policy documents does not take into account the actual implementation. Thus, implicit or explicit gender considerations in the EU's policy practice lie beyond the scope of this article.
The analysis proceeded in three steps First, I downloaded all documents the Commission lists as 'legislative documents' of the Agenda on Migration on its website (European Commission, 2019) and identified those that are relevant to the analysis. An automated lexical search for key words like gender, women/men, female/male, sexuality, and LGBT/LGBTI helped to identify policy areas that referenced gender. Table A1 in the online appendix summarizes all analysed documents.
Second, assisted by the programme MaxQDA, I read and manually coded the documents listed in Table A1 through the lens of CFA's synthesizing questions. The questions were adapted to the context at hand, for example, 'what is considered to cause the crisis' and 'what does gender mean in this context?' Third, I grouped the documents into internal and external policy instruments and looked for similarities and differences with regard to how they refer to gender. Before I present the analysis of policy responses, the following section discusses how the Agenda on Migration represents the crisis.

Defining a Multifaceted Crisis: The EU's Agenda on Migration
In general, the EU's commitment to gender equality in the field of migration and asylum results from its general subscription to gender norms in EU and international law, for example, its commitment to Gender Mainstreaming in the Treaty of Amsterdam. In addition, the Convention on the Elimination of all Forms of Discrimination against Women and the Council of Europe's 2011 Istanbul Convention on preventing and combating violence against women and domestic violence should also apply to refugees. The first demand for more gender sensitivity in the field of refugee policies came from the European Parliament. Specifically, in its Resolution of 13 April 1984, it demanded a gender-sensitive reading of the Geneva Refugee Convention of 1951. Yet, only with the 1999 Treaty of Amsterdam were EU institutions granted power to develop legislation in the area of asylum. The European Council's adoption of the Tampere Programme of October 1999 envisaged the development of common asylum standards and procedures. Until 2006, the EU developed the Asylum Procedure Directive, the Reception Conditions Directive, the Qualification Directive, and the Dublin and the EURODAC regulation. Despite some limitations, gender and sexual orientation were increasingly incorporated into the CEAS (AIDA, 2017; Tsourdi, 2011), while references to gender in EU external refugee policies remained limited.
The 2015 increase in refugee arrivals to the EU and the fatalities in the Mediterranean created significant pressure to develop a coordinated crisis response. The EU's Agenda on Migration serves as a guiding document for the EU's reaction to the crisis and is therefore central to understand the underlying problem representation. The Agenda sees its raison d'être in the need to restore confidence in our ability to bring together European and national efforts to address migration, to meet our international and ethical obligations and to work together in an effective way, in accordance with the principles of solidarity and shared responsibility (European Commission, 2015a).
To address this multifaceted problem definition, the Agenda reframes existing policies and introduces new ones. Immediate actions consist of key emergency responses like saving lives at sea, fighting smugglers, relocation, resettlement, cooperation with third countries and support for frontline Member States like Italy or Greece. The EU's long-term vision 'to manage migration better' comprises four pillars: (1) 'reducing incentives for irregular migration', (2) 'border management: saving lives and securing borders', (3) 'Europe's duty to protect: a strong common asylum policy' and (4) 'a new policy on legal migration' (European Commission, 2015a: 3-6). Oriented towards the external, the first two pillars comprise cooperation on border control with countries of origin and first refuge, development cooperation, fighting smuggling activities and trafficking in human beings, return and increasing the activities of FRONTEX.
Overall, all externally oriented policy proposals aim to prevent mobility towards the EU, regardless of individual or group-specific protection needs. Targeting development cooperation towards the 'fight of root causes' of migration and expanding FRONTEX activities against smugglers and trafficking, the EU adopts a security-oriented and militarized crisis response (see Muehlenhoff and Hoijtink, this issue). Despite the declared goal to 'save lives' (European Commission, 2015a: 3), FRONTEX operations like Triton, Poseidon and Sophia are military naval missions, whose principal mandate is the 'disruption of the business model of human smuggling and trafficking networks' (European Council, 2015). There is no mention of how gender in interaction with other social categories shapes vulnerabilities throughout different phases of the migratory process. While underscoring different motivations for fleeing to Europe (European Commission, 2015a: 2, 7), the Agenda conceals refugees' individual, diverse and complex situations, subsuming them all under the term of 'migration'. Migration then becomes an agent and actorfree problem that the EU can 'tackle', 'fight' and 'manage'. This depersonalized and depoliticized representation of refugees disregards the heterogeneity of this group, their individual experiences and the deeply political dimensions of their flight. As Horst (2006: 14) puts it, 'the complex identities of a single person are taken and replaced by one: that of the refugee. [. . .] Refugees are stripped of specific features of their society, their place of origin and history'.
The only social category singled out as vulnerable in the Agenda are 'children', whose fundamental rights should be protected by monitoring reception and asylum procedure standards (European Commission, 2015a: 12). However, children's special protection needs become only recognized within the EU. Thereby, the EU border marks a geographical line from where on refugee children's rights and protection needs start to matter, while the relevance outside the EU is disregarded. Neither this passage nor the rest of the document refer explicitly to gender, sexuality or any other social categories at any time.
As a whole, President Juncker's Agenda of Migration represents the crisis as a problem of numbers, migratory pressure and legal versus irregular migration. Internally, it frames EU action in terms of the protection of refugees, and externally as protection of EU borders. The absence of an explicit commitment to gender norms in the EU's Agenda on Migration persists in most legislative texts, however with some exceptions.

Gender in EU Internal Policy Responses
Internal responses to the crisis were mainly concerned with emergency measures for Southern EU Member States, EU internal distribution of refugees via relocation and EU internal protection standards under the CEAS (cf. Table A1). Due to their geographical proximity to refugee hosting states along with existing EU visa regulation and the Dublin system, Southern Member States have faced significantly higher numbers of arrivals when compared to other Member States. One response to this was the so-called 'hotspot-approach'. According to the EU, these 'first reception facilities' aim 'to better coordinate EU agencies' and national authorities' efforts at the external borders of the EU, on initial reception, identification, registration and fingerprinting of asylum-seekers and migrants' (Mentzelopoulou and Luyten, 2018: 1). The report on the EU's implementation of the hotspot approach (European Commission, 2017b: 6) designates unaccompanied minors, victims of trafficking as well as 'shipwreck victims, single women, victims of violence etc.' as particularly vulnerable groups. The Commission requests that Member States apply a 'minimum gender balance among the deployed experts' and have 'safe and separated areas' for vulnerable persons (European Commission, 2017b: 3).
A second and related policy proposition to support Southern Member States was the EU relocation mechanism. Established in 2015, it aimed to relocate 106,000 persons in need of international protection. In November 2017, EU Migration Commissioner Avramopoulos declared that the implemented facility was to be closed with around 31,500 persons relocated from inauguration until time of closure (European Commission, 2017d). The relocation mechanism gave priority to 'vulnerable applicants within the meaning of Article 21 and 22' of the reception Recast Directive (European Commission, 2015b: 16). It defines 'vulnerable persons' as, inter alia, 'minors, unaccompanied minors, disabled people, elderly people, pregnant women, single parents with minor children, victims of human trafficking, persons with serious illnesses, persons with mental disorders and persons who have been subjugated to torture, rape or other serious forms of psychological, physical or sexual violence, such as victims of female genital mutilation (European Union, 2013). Thereby, in contrast to EU external policies' reductionist definitions, the directive offers a comprehensive understanding of vulnerabilities. Special protection needs are not limited to children and women per se, but also to the elderly, to single parents irrespective of their gender, and to different forms of bodily and medical conditions, such as female genital mutilation. Yet, sexual orientation, for instance, does not figure in the list.
To address the internal dimension of varying standards and procedures, the EU proposed another reform to the CEAS, amending the EU legal norms on qualification, reception and procedures (European Commission, 2016b, 2016d. Like in the previous version of the CEAS, the proposals demonstrate a clear commitment to gender sensitivity and further extend the recognition of gender. To illustrate, the proposal for a common procedure for international protection (European Commission, 2016d) points to the Charter of Fundamental Rights and to the right to non-discrimination, equality of rights between men and women, and the rights of the child. It invokes Member States' obligations under the Istanbul Convention on preventing and combating violence against women and domestic violence. Provisions aim to ensure 'substantive equality between female and male applicants' through same-sex interpreters, medical examiners and the opportunity for spouses to be interviewed separately (European Commission, 2016d). The proposal for a revised reception directive (European Commission, 2016b) suggests several changes with regard to gender, for example, that: the specific needs of women applicants who have experienced gender-based harm should be taken into account, including via ensuring access, at different stages of the asylum procedure, to medical care, legal support and to appropriate trauma counselling and psycho-social care.
Furthermore, Article 24 adds gender-based harm to the definition of 'victims of torture and violence' (European Commission, 2016b: 26, 61).
All factors considered, some internally oriented EU responses to the crisis demonstrate a comprehensive understanding of gendered vulnerabilities and include concrete propositions of how to address them. Internally, the EU emphasizes refugees' diverse protection needs and remains committed to upholding international human rights.

Gender in EU External Policy Responses
External policy areas with explicit references to gender are the EU funding schemes, the EU's cooperation on migration with third countries, especially with Turkey, and the EU proposition for an EU resettlement framework. In addition, a joint communication on 'Migration on the Central Mediterranean route' identifies the increasing number of 'vulnerable migrants, especially women and minors', crossing the Mediterranean as a 'worrying trend' (European Commission, 2017c: 4). By emphasizing the deaths of women and children in the Mediterranean, the communication creates an implicit gendered and agedifferentiated hierarchy of suffering. This reinforces stereotypical gender notions in forced migration, claiming that '"authentic" refugees are women and children, who are implicitly vulnerable and in need of external assistance' (Turner, 2017: 29), while male refugees' vulnerabilities are disregarded. Such gendered representations also resonate with male refugees' portrayal as potentially dangerous and undeserving of protection in broader societal and public debates (Rettberg and Gajjala, 2016;Vollmer and Karakayali, 2018).
Along a similar line, the EU's funding schemes, aimed at fighting 'root causes' of migration, mention gender only in relation to the 'special needs' of women and children. EU funds like the Trust Fund for Africa or the Asylum, Migration and Integration (AMIF) fund have been used to address conditions in refugees' regions of origin. These funds equate gender with women and consistently group the former with children and youth. For instance, various EU development projects under the AMIF funding scheme consider women, children and youth as the target group for their actions and combine child-related activities and actions to prevent sexual and gender-based violence. In these mostly development-oriented external responses, gender considerations result in concrete policy actions. Yet, this grouping of women and children under a victimization frame infantilizes female refugees and ignores the agency that both children and refugee women can have in the migratory process. The one-dimensional categorization of vulnerability ignores the complex intersections of gender or age with other social categories like ethnicity, religion, social class and the body. Moreover, this categorization implicitly disregards refugee men and their experience. Reorienting EU development cooperation towards the fight of 'root causes' limits gender to the categories of vulnerable women and children, whose protection needs the EU seeks to address in their regions of origin.
Besides development instruments, the EU-Turkey Statement of 18 March 2016 constitutes one of the core external policy responses to the crisis (European Council, 2016). It details Turkey's commitment to stricter border control on both land and sea. In exchange, the EU promises a voluntary humanitarian admission scheme, visa liberalizations for Turkish nationals, financial assistance for projects benefitting Syrian refugees in Turkey as well as upgrading the Customs Union and the recommencement of the accession process (European Council, 2016). In regular Progress Reports, the EU Commission communicates the statement's implementation. With its focus on reducing the number of arrivals to the EU, the statement aligns with the EU's crisis definition: disregarding the question of how many refugees in Turkey are in need of assistance. Overall, the EU-Turkey statement has no explicit commitment to a gender-sensitive approach or gender promotion in Turkey.
Nevertheless, the allocation of humanitarian assistance under the Facility for Refugees in Turkey refers explicitly to gender in certain sections. The first progress report of 20 April 2016 lists 'child protection, women's health, and education in emergencies' as examples of vulnerable groups that shall benefit from EU humanitarian assistance in Turkey (European Commission, 2016a). The sixth progress report (European Commission, 2017f: 11-12) stresses that 'in its delivery, the Facility for Refugees in Turkey supports rights of children, and human rights of refugees in general, including gender equality' as well as 'vulnerable refugees including victims of gender-based violence'. The subsequent report (European Commission, 2017e: 11) declares 'particular attention to the situation of human rights of refugees overall and support [for] in particular women, children and disabled persons'. At least on paper, considerations of gendered vulnerabilities seem to gain importance over time.
A concrete example of gender considerations can be found in the Emergency Social Safety Net. The EU Commission launched the facility together with Turkish authorities, aiming to progressively cover 'one million of the most vulnerable refugees' by the first half of 2017. The EU allocated humanitarian assistance to projects that aim at preventing sexual and gender-based violence in South Turkey, and to protecting survivors of genderbased violence. Yet, only 3 out of 57 EU projects under the Facility for Refugees in Turkey explicitly mention a gender-related focus, amounting to only €18,000,000 out of 2.9 billion (European Commission, 2017a).
A second dimension of gender in the EU-Turkey statement can be found in the EU's progress assessment of the visa liberalization roadmap. The Commission Staff Working Document (European Commission, 2016f), accompanying the third report on 'progress by Turkey in fulfilling the requirements of its visa liberalization roadmap', lists 72 requirements of this roadmap and describes Turkey's implementation progress. The overall problem representation remains irregular migration to the EU, for which Turkey's improved 'border management' is presented as the solution. The purely number-driven problematization of Turkey's borders stands in sharp contrast with a normative set of requirements in the chapter on fundamental rights. It requires Turkey to ensure the freedom of movement of citizens without discrimination based on any ground such as 'sex, race, colour, ethnic or social origin, genetic features, language, religion or belief, political or any other opinion, membership of a national minority, property, birth, disability, age or sexual orientation' (European Commission, 2016f: 24). The following requirement requests Turkey to ensure 'full and effective access to travel and identity documents for all citizens including women, children, people with disabilities, persons belonging to minorities, internally displaced people, and other vulnerable groups' (European Commission, 2016f: 35-36). By instructing Turkey to consider gender and other discrimination grounds, the EU stages itself as the global teacher of gender and non-discrimination norms. Thereby, the EU further externalizes not only its border control but also responsibility for realizing normative commitments, such as gender equality, to third countries.
The Proposal for a Union Resettlement Framework (European Commission, 2016c) also includes several explicit references to gender. Resettlement aims to show solidarity with countries of first refuge by granting a limited number of refugees safe and legal access to a third country that has agreed to admit them (United Nations High Commissioner for Refugees (UNHCR), 2011: 3). The proposal's section on fundamental rights refers to the Istanbul Convention and the UN Convention on the Rights of the Child and stresses that 'a gender-sensitive approach should be adopted when interpreting and applying this Regulation'. Furthermore, the proposal subscribes to the general principles of equality and non-discrimination (European Commission, 2016c: 8-9). Echoing to a large extent the resettlement criteria of the UNHCR, the proposal lists, inter alia, 'women and girls at risk' and 'survivors of violence and/or torture, including on the basis of gender' as eligible groups for resettlement to the EU. While the focus on refugee women can also have exclusionary effects for single, young men (Turner, 2016), the reference to gender sensitivity is nevertheless striking. The fact that the framework comes in the form of a regulation, which -in contrast to ad hoc 'deals' like the EU-Turkey statement -has to pass EU policy-making procedures with more influence of norm advocates, may explain this outlier among EU external policies.

Conclusion
This article departed from the question of to which extent the crisis offered a window of opportunity for the EU to expand its considerations of refugees' gendered vulnerabilities. Empirically, the analysis demonstrated a divide between EU-internally and EU-externally oriented crisis responses. Internally, definitions of vulnerability avoid an exclusive focus on female and minor refugees, and instead endorse a wide range of social markers that can constitute, in interaction with gender or in and of themselves, special protection needs. Referring to the EU's commitments to gender equality and human rights, gender considerations take the form of concrete provisions and actions towards more gender-sensitive refugee protection.
In EU external crisis responses, gender mostly refers to the category of vulnerable women and children, whose protection needs the EU seeks to address beyond EU borders in countries of origin or first refuge. To this end, EU development cooperation -a policy field with a comparatively comprehensive gender equality agenda before the crisis -is reoriented towards 'fighting root causes' of migration. As Allwood (2015Allwood ( , 2019 has shown, this migration-development nexus significantly reduces gender considerations. Another way EU external responses address gendered vulnerabilities is through making gender considerations a conditionality of benefits for third countries, for instance, through visa liberalizations for Turkey. Such policies frame the EU as a teacher of gender and non-discrimination norms towards third countries while at the same time containing refugees' mobility towards the EU. Thus, similar to the way the EU intertwines security and humanitarian dynamics in the Mediterranean (Moreno-Lax, 2018;Pallister-Wilkins, 2018;Perkowski, 2018), gender considerations in EU external crisis responses paradoxically work in the benefit of border, not refugee, protection.
The juxtaposition of gender in internal and external crisis responses deepens our understanding of how the EU's normative commitments to gender equality develop in times of crisis. Crisis transforms the very meaning and scope of gender considerations by aligning them, wherever needed, with what is considered to be the solution to the crisis. Externally, the focus on reducing arrivals to the EU kept gender considerations limited, while internally, the demand for further harmonizing asylum standards and procedures allowed the expansion of already existing gender considerations. This paradoxical relation between gender in external and internal crisis responses, which marginalizes the EU's concern for gendered vulnerabilities beyond EU borders, makes an uncritical appraisal of EU internal progress challenging.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research is funded by a Research Talent Grant (Project Number 406-16-535) by the Netherlands Organisation for Scientific Research.

Supplemental Material
Additional supplemental material may be found with the online version of this article.  (Baerwaldt, 2018;Gilbert, 2015). Niemann and Zaun (2017) propose to call it a crisis of the Common European Asylum System (CEAS), to underscore that not refugees themselves caused the crisis but rather the shortcomings of the CEAS. For the convenience of the reader, I will only speak of 'crisis' in the following. 2 The EU usually refers to migration without specifying it as forced or voluntary. Scholarship uses the term 'forced migration' to delineate studies that are interested in the specific situation of or policies for people who are internally displaced or had to flee their home country due to individual persecution, war and violence (Castles, 2003). Analytically, the distinction is important for the context at hand because gender plays out differently in forced and voluntary migration.