Australian Year Book of International Law
Women’s rights are at the forefront of developments in international refugee law. A refugee is a person who is unable or unwilling to avail himself or herself of the protection of his or her country, owing to a well-founded fear of persecution for reasons of race, religion, nationality, membership of a particular social group or political opinion. The archetypal refugee is a person who has fled his or her country in order to avoid being tortured by state officials on account of his or her political opinions. However, the outer limits of the refugee definition have been challenged and expanded by a number of cases in Australia over the past decade. These cases have focused on issues ranging from persecution by non-state actors to the meaning of a ‘particular social group’. One of the most significant challenges to the traditional interpretation of the refugee definition has come from a gender perspective. While the refugee definition appears to apply equally to men and women, in practice women have not benefited equitably from refugee protection. This inequality derives, in part, from the refugee definition having traditionally been interpreted by reference to the male-dominated public sphere.
This article examines the recent ground-breaking case of the Minister for Immigration and Multicultural Affairs v Khawar, which concerned the persecution of a woman in the private sphere. This case provides an excellent vehicle through which to explore the issues of gender and public/private distinctions more generally in refugee law. Khawar, who was a Pakistani citizen, claimed that she was a victim of domestic violence. According to her application, she approached the police in Pakistan for help on four occasions and each time they either refused to take her complaint or failed to take it seriously. Khawar alleged that the police in Pakistan systematically discriminated against women by failing to protect them in domestic situations and that this discrimination was tolerated and sanctioned by the state. Thus Khawar argued that she was a refugee under the Convention definition because she had a well-founded fear of persecution, in the form of lack of state protection against domestic violence, and that this persecution was for reason of her membership of a particular social group, such as ‘women in Pakistan’.
The High Court held that the failure by a state to provide effective protection to women from domestic violence could amount to persecution for a Convention reason. This decision is significant because it helps to bridge a gendered public/private divide in international refugee law by recognising that a state has a public duty to protect its citizens from private harms. Part I of this article provides a theoretical overview of the relationship between gender and public/private distinctions in international law and refugee law. Part II outlines the major cases dealing with the meaning of ‘persecution’ and ‘particular social group’ in other jurisdictions. Part III sets out the history and judgment in Khawar, while Part IV provides a critical analysis of the decision. Part V explores some of the continuing challenges posed by a gender analysis of refugee law, which either build upon or qualify the decision in Khawar. The High Court’s decision in Khawar represents an important step towards interpreting the refugee definition in a more gender sensitive way to incorporate the experiences of women. However, the factual basis of the case and the decision evidence limitations that leave room for further developments in future cases.
There is no single public/private distinction in legal discourses. Rather, public/private distinctions are multi-layered and multi-faceted. What is public in one context may be private in another and the content of these categories may also vary over time. For example, public/private distinctions are used to distinguish between public (state) and private (non-state) actors, as well as public (political) and private (market) issues. They are also used to distinguish between the public sphere (including politics and the market place) and the private sphere (including the home and family). While public/private distinctions can be analysed from a variety of perspectives, this article examines such distinctions by reference to divisions based on gender.[7 ] Although this only reflects one aspect of public/private distinctions, it provides an important way of identifying the gendered nature of law.
In Western liberal states, the law has traditionally regulated activities that occur in the public sphere, including state activity and the market place, but not those that occur in the private sphere, including family life and personal activities. O’Donovan defines the private sphere as the area of life into which the law will not intrude or which is not the law’s business. The phrase ‘a man’s home is his castle’ encapsulates the liberal notion that intervention by the state in the private sphere of family life was to be avoided if possible. This division is problematic because women have traditionally been confined to the private sphere, while men have been able to occupy the public arena as well. Non-intervention by the state within the private sphere has meant that women often lack rights and protection where they are most vulnerable. For example, violence perpetrated within the private sphere, such as domestic violence, has typically been regarded as being less important than violence perpetrated in the public sphere. For this reason, feminist scholars have challenged the notion of the family as a sacrosanct private sphere with maxims such as ‘the personal is political’.
International law rests on and reproduces similar gendered public/private distinctions. International law has traditionally been limited to the public sphere of relations between states, rather than encompassing the private sphere of relations within states. Just as the police were reluctant to intervene in the private sphere of the home, international law has been reluctant to intrude into the private sphere of domestic law enforcement. This public/private distinction has been challenged by a number of developments over the last 50 years, in particular by human rights law. International human rights law crosses the traditional boundary between international and domestic law because it deals with intrastate issues, such as how states treat their own citizens. However, human rights law still operates almost exclusively within the domestic public sphere, as civil and political rights generally protect against governmental action, while economic and social rights concern rights such as the right to work in the paid workforce. As women and women’s issues have traditionally been relegated to the private sphere domestically, women’s experiences and concerns are still marginalised in international human rights law.
A factor that has contributed to human rights law being largely confined to the public sphere is that violations of human rights often require some form of state action. For example, consider the international prohibition on torture. Although the prohibition is meant to protect the ‘inherent dignity of the human person’, the Convention Against Torture defines torture as severe pain or suffering inflicted by, or at the instigation of, or with the consent or acquiescence of, a public official or person acting in an official capacity. Thus torture tends to be limited to acts occurring in the public sphere because the state must be implicated in the act. Pain or suffering inflicted within the private sphere that has no direct connection to the state, such as domestic violence, does not qualify as torture even if it undermines the inherent dignity of the human person. Instead, private harms are left to domestic law enforcement, even where such enforcement is inadequate. Thus international human rights law has traditionally excluded some of the most pervasive forms of harm against women. According to MacKinnon, ‘when men use their liberties socially to deprive women of theirs, it does not look like a human rights violation. But when men are deprived of theirs by the government, it does.’
Even though there are treaty articles providing for equality between men and women, women’s rights have not been given the same level of protection or attention as other human rights.[20 ] This failure represents a striking blind spot in the early development of the human rights movement. However, this deficiency has been strongly challenged by feminist commentators, particularly over the last decade, who have argued that ‘women’s rights are human rights’. This has led to progress in a number of areas, including increased attention to violence against women. For example, in 1989, the CEDAW Committee recommended that states parties include information about violence against women, and measures taken to counter that violence, in their periodical reports. In 1992, the CEDAW Committee adopted a General Recommendation stating that gender-based violence is a form of discrimination under the Convention, whether perpetrated by a person, organization or enterprise. In 1993, the United Nations General Assembly adopted a Declaration on the Elimination of Violence Against Women, which specifically addressed violence within families, dowry-related violence and female genital mutilation. In 1994, Coomaraswamy was appointed as the first special rapporteur on violence against women, its causes and consequences. Thus, international human rights law has begun to bridge some of its gendered public/private distinctions.
The operation of gendered public/private distinctions in international law provides an important backdrop for understanding the traditional reluctance of refugee law to accept private harms as the basis for refugee status, as well as for appreciating the modern developments challenging this conception. The central principle of the Refugee Convention is non-refoulement, which requires states not to expel or return refugees to the frontiers of territories where their life or freedom would be threatened on account of their race, religion, nationality, membership of a particular social group or political opinion. The definition of ‘refugee’ is critical because it determines whether or not protection duties are owed to a particular individual. The Convention makes no formal distinction between male and female refugees, so it appears to apply equally to both. However, the definition has primarily been interpreted by reference to the male-dominated public sphere. This has served to marginalise the experiences of women because women often experience harm in the private sphere by non-state actors or ‘mere discrimination’ rather than persecution in the public sphere.
Refugee law has traditionally dismissed gender-specific harms and gender-related persecution as private harms occurring for personal reasons. The meaning given to ‘gender-specific harm’ and ‘gender-related persecution’ differs between commentators. I use the phrase ‘gender-specific harm’ to refer to harm that is unique to, or more commonly befalls, members of one sex. This includes sexual violence, genital mutilation, forced marriage, forced abortion and domestic violence. By contrast, ‘gender-related persecution’ refers to a causal relationship between the persecution and the reason for the persecution. Gender-related persecution means that a woman is persecuted because she is a woman, whether or not she is persecuted in a gender-specific way. Gender-specific harms and gender-related persecution may be, but are not necessarily, linked. A woman may be persecuted using gender-specific means for reasons unrelated to gender (eg being raped due to her membership of a particular race or political party), or she may be persecuted using non-gender-specific means but for gender-related reasons (eg being flogged for refusing to wear a veil) or she may be persecuted using gender-specific means for gender-related reasons (eg female genital mutilation).
Women asylum seekers who fear harm in the private sphere, such as domestic violence, face three main problems in establishing their refugee status. First, they must prove that the harm they apprehend is sufficiently serious to amount to persecution. Not all violations of human rights amount to persecution. According to the United Nations High Commissioner for Refugees (UNHCR) Handbook on Procedures and Criteria for Determining Refugee Status, persecution includes a threat to life or liberty, as well as other serious violations of human rights, for one of the Refugee Convention reasons. Persecution has also been defined as the ‘sustained or systematic violation of basic human rights demonstrative of a failure of state protection’. While persecution involves systematic and discriminatory conduct, only discrimination that leads to substantially prejudicial consequences will amount to persecution. Individual acts that might not amount to persecution when examined in isolation, might amount to persecution when viewed together.
Gender-specific harm has sometimes been trivialised as a private matter that does not amount to persecution. For example, rape in wartime has traditionally been considered to be a private misfortune rather than a public harm. However, rape and other forms of sexual violence are now recognised as serious harms that can amount to torture, or to cruel, inhuman or degrading treatment. Similarly, other forms of harm, such as domestic violence, have traditionally been viewed as private harms that could not found a claim to refugee status . The widespread occurrence and cultural acceptance of domestic violence have also been cited as reasons against characterising such violence as persecution, even though these factors actually make the harm feared more serious because the police are more likely to turn a blind eye to such harms. Women are also likely to suffer discrimination in achieving equality of rights, in particular economic and social rights, such as the right to work, to travel and to be educated. These rights have traditionally been characterised as less important in refugee law than civil and political rights, which can have gendered effects. Thus discrimination against women, such as the imposition of strict dress codes, has been dismissed as mere discrimination, when it is arguable that it may constitute persecution.
Second, the state must be implicated in the infliction of the harm. In the paradigm case, the persecution feared is persecution by the state or its agents for a Refugee Convention reason. In these cases, the persecution involves direct violations of human rights in the public sphere by the state. However, many of the serious harms perpetrated against women take place in the private sphere by non-state actors where the state does nothing to prevent the harm. In these cases, the persecution involves state inaction rather than state action. For example, domestic violence and female genital mutilation often take place at the hands of women’s families or communities, where the state does not take adequate steps to protect women from these harms. According to Wright, for ‘most women, indirect subjection to the state will almost always be mediated through direct subjection to individual men or groups of men.’ However, the UNHCR Handbook provides that while persecution is normally related to actions by the state, serious discriminatory or other offensive acts committed by non-state actors ‘can be considered persecution if they are knowingly tolerated by the authorities, or if the authorities refuse, or prove unable, to offer effective protection’. Thus states have a duty not to persecute and a duty to provide protection against persecution by non-state actors. However, as the paradigm case concerns state actors rather than non-state actors, and state action rather than state inaction, courts have often been reluctant to find persecution where states have systematically failed to provide effective protection to women in the private sphere. This results from confusion between the archetypal example of a refugee and the Convention definition of a refugee.
Finally, the persecution must be for reasons of race, religion, nationality, membership of a particular social group or political opinion. Sex and gender are not specifically included so harm that is motivated by gender-related considerations is not expressly covered by the Refugee Convention. Some commentators have argued that gender should be added as a separate Convention ground, while others have favoured reinterpreting the existing Convention grounds to include the experiences of women. At present, women asylum seekers who fear such harm must show that they fear persecution for a Convention reason such as their religion, political opinion or membership of a particular social group. The problem with this approach is that these Convention grounds have been given incomplete and gendered interpretations. For example, the traditional exclusion of women and women’s issues from the public sphere means that political protests and resistance by women often manifest themselves in different ways to men. While overt expression of a political opinion through conventional means (such as involvement in a political party) has been considered a basis for asylum, less-conventional forms of political resistance (such as women failing to abide by discriminatory laws or codes of conduct) have often been categorised as personal conduct.
Courts have traditionally dismissed gender-specific harms, such as rape and domestic violence, as private crimes perpetrated for personal reasons. Even victims of rape by military personnel during times of war have found it difficult to establish their refugee status. Courts have tended to focus on the reason why the non-state actor perpetrated the harm, rather than on the reason why the state systematically failed to protect against that harm. There has also been insufficient attention given to whether women are persecuted because they are members of a particular social group, such as women, or for other Refugee Convention reasons, such as political opinion or religion. For example, a woman who objects to laws or religious practices that discriminate against women may suffer discrimination on account of her feminist political opinions or religious beliefs.
Just as there has been a growing awareness that women’s rights are human rights, there has also been a rising consciousness of gender issues in refugee law. Recognition of gender-bias in both the refugee definition and the process of refugee determination has led to the adoption of gender guidelines for refugee decision-makers by the UNHCR, Canada, the United States and the United Kingdom. In 1996, the Australian Department of Immigration and Multicultural Affairs similarly issued Guidelines on Gender Issues for Decision-makers. While these administrative guidelines do not have legal force, they provide assistance to initial decision-makers about how to interpret the refugee definition in a gender sensitive manner. Apart from these developments, there have been a number of cases in other jurisdictions dealing with the meanings of ‘persecution’ and ‘particular social group’, which are outlined below in chronological order.
The leading Canadian case on the meaning of ‘particular social group’ is Canada (Attorney-General) v Ward. Ward had been a member of the Irish National Liberation Army but had since been placed on a death list by that group. Justice La Forest, who delivered the judgment of the Court, noted that the other categories of race, religion, nationality and political opinion also featured prominently in human rights law. He thus found inspiration in determining the meaning of ‘particular social group’ by reference to human rights and anti-discrimination concepts. Accordingly, he held that a social group includes three types of groups. The first group is defined by an innate or unchangeable characteristic of its members, such as gender, linguistic background and sexual orientation. The second group includes members who voluntarily associate for reasons so fundamental to their human dignity that they should not be forced to forsake the association, such as human rights activists. However, in a later case, La Forest J clarified that membership of such a group need not be voluntary. Members of the third type of group are defined by a former voluntary status, which is unalterable due to its historical permanence. This category recognises that a person’s past is an immutable part of that person.
In Mayers and Marcel v Minister of Employment and Immigration (Canada), the Canadian Federal Court of Appeal considered whether ‘women’, ‘Trinidadian women’ or ‘Trinidadian women subject to wife abuse’ could constitute a particular social group. The Court found, obiter dictum, that the word ‘particular’ modifies the meaning of ‘social group’ because it suggests some limits on the category. It may be that ‘women’, constituting half of humanity, could not aptly be described as a particular social group. The same may be true of ‘Trinidadian women’. By contrast, there was evidence on which a tribunal might find that ‘Trinidadian women subject to wife abuse’ were a particular social group. However, Mahoney JA expressed concern about whether fear of persecution could be the only factor that distinguishes a particular social group from a social group at large.
There have been two main interpretations of the meaning of ‘particular social group’ in the United States. In In re Acosta, the Board of Immigration Appeals (the Board) applied the maxim of ejusdem generis to construe the general phrase ‘particular social group’ by reference to the genus of the other categories listed. The Board found that race, religion, nationality and political opinion referred to immutable characteristics that were either beyond the power of an individual to change (eg, race or nationality) or so fundamental to individual identity or conscience that they ought not be required to be changed (eg, religion or political opinion). These immutable characteristics may be innate, such as sex, colour or kinship ties.
By contrast, in Sanchez-Trujillo v Immigration and Naturalisation Service, the Ninth Circuit Court of Appeals rejected a claim that a social group existed of young, urban, working-class males of military age who had not served in the military. The Court rejected the conception of a particular social group as involving sweeping demographic divisions and instead found that the phrase implied a collection of people closely affiliated with each other and actuated by some common impulse or interest. This decision has been widely criticised as being unduly restrictive, in particular because of the Court’s emphasis on voluntary relationships. However, more recently in Hernandez-Montiel v Immigration and Naturalisation Service, the Ninth Circuit Court of Appeals noted that other circuits had generally adopted the test from In re Acosta, while it had departed from that test in Sanchez-Trujillo. The Court attempted to reconcile the different tests by the expansive statement that a particular social group ‘is one united by a voluntary association’ or by an ‘innate characteristic that is so fundamental to the identities or consciences of its members that members either cannot or should not be required to change it’.
There have also been a number of cases dealing with persecution of women by non-state actors. In In re Kasinga, the Board granted asylum to a young Togolese woman who feared female genital mutilation where the state had taken no steps to prevent the practice. The relevant social group was defined as young women of the particular tribe, who had not had genital mutilation and who opposed the practice. Applying In re Acosta, the Board found that some of these characteristics were innate (such as being a young woman) while others were so fundamental to the claimant’s individual identity that she should not be required to change them (such as having intact genitalia). In Fatin v Immigration and Naturalisation Service, the female asylum seeker was opposed to Iranian fundamentalist laws and oppressive gendered practices. The Third Circuit Court of Appeals found that women could constitute a particular social group, but that the claimant had only submitted evidence of women being persecuted if they refused to obey the gendered laws, rather than evidence of persecution of women generally in Iran. With respect to cases involving domestic violence, in In re R-A-, the Board did not grant asylum to a Guatemalan woman who had suffered severe domestic violence because the violence was characterised as a private act. However, this case attracted considerable criticism and in January 2001, the US Attorney-General issued an order vacating the decision and remanding the matter to the Board for reconsideration. In In re S-A-, the Board granted asylum to a domestic violence victim where religious fundamentalism also played a pivotal role in the persecution.
The leading case on this issue in the United Kingdom is the joint decision of Islam v Secretary of State for the Home Department; R v Immigration Appeal Tribunal; Ex parte Shah. Islam and Shah were both citizens of Pakistan whose husbands had falsely accused them of adultery and been violent towards them. They were abandoned by their husbands and lacked the protection of other male relatives. If returned to Pakistan, they risked criminal proceedings for sexual immorality under Sharia law, which carried penalties including flogging and stoning to death.
The House of Lords accepted, by a majority of four to one, that the husbands’ violence combined with a lack of state protection amounted to persecution for reasons of the women’s membership of a particular social group. Three of the four judges in the majority, Lords Steyn, Hoffmann and Hope of Craighead, agreed that ‘women in Pakistan’ were a particular social group. Lord Steyn would also have accepted that the applicants were members of a more narrowly defined group based on their gender, the suspicion of adultery and their unprotected position in Pakistan. Lord Hutton agreed that the applicants were members of this narrower social group, but did not express an opinion on the broader social group.
Lord Steyn noted that while domestic violence against women was prevalent in many countries, the distinguishing feature was that, on the unchallenged evidence of the case, domestic violence was tolerated and sanctioned by the state of Pakistan. Lord Hoffmann found that the persecution was made up of two elements: the threat of violence by each husband, which was a personal affair; and the inability or unwillingness of the state to do anything to protect the women, which was because they were women. Lord Millett, in dissent, found that the narrower social group, of married women in Pakistan who were under a suspicion of adultery and were unprotected, did not exist independently of the persecution. As to the broader social group of ‘women in Pakistan’, he found that the applicants were not persecuted because they were women. Instead, they were persecuted because they had transgressed social norms and there was no evidence that men who transgressed social norms that applied to them were treated more favourably.
In Refugee Appeal No 71427/99, the Refugee Status Appeals Authority considered an asylum claim made by an Iranian woman who had suffered domestic violence at the hands of her ex-husband. Haines QC found that women had unequal rights in marriage, divorce and custody issues in Iran. With respect to domestic violence, he found that the state of Iran condoned, if not actively encouraged, non-state actors such as husbands or former husbands to cause serious harm to women. Haines QC concluded that the claimant had suffered persecution for reasons of religion and her political opinions, as Iran is an Islamic theocratic state. However, he found that the overarching reason why the applicant had a well-founded fear of persecution was that she was a woman and that women constituted a particular social group.
These decisions from other jurisdictions provide a good background for understanding the Australian approach to these issues in Khawar.
In Khawar’s application for refugee status, she claimed that she had been the victim of domestic violence by her husband and his family for more than 10 years. This abuse included slappings, beatings that in some cases led to hospitalisation, threats to throw acid on her and threats to kill her by setting her on fire. On one occasion, her husband and his brother poured petrol on Khawar and threatened to set her alight; they only desisted when a neighbour arrived in response to Khawar’s screams. Against this evidence, there was anonymous information before the Tribunal that Khawar’s claims of domestic violence were false and that her husband had followed her to Australia and that they were on good terms.
Khawar claimed that on four occasions she had approached the police in Pakistan to complain about domestic violence and they had not offered her any protection. On the first occasion, the police officer refused to take the complaint and told her that such incidents were occurring throughout the country and the police could do nothing about it. On the second occasion, Khawar went to the police with her sister’s husband, believing that she might be taken more seriously if accompanied by a man, but the police officer did not take down her story accurately and did not seem to take her seriously. On the third occasion, Khawar again took her sister’s husband and the same police officer wrote an inaccurate report and offered her no assistance. On the fourth visit, the police officer told her that women always tried to blame their husbands for problems which they themselves caused and that she should go and do her own work.
Khawar’s solicitor also filed material concerning the position of women in Pakistan, which included reports from the United States State Department, the Canadian Immigration and Refugee Board, Amnesty International and the Australian Department of Foreign Affairs and Trade. The effect of this material was that women were not treated as equal citizens in Pakistan. Some laws in Pakistan were discriminatory on their face, for example, laws that devalued the testimony of female witnesses compared with male witnesses. Other laws were applied in a discriminatory manner, for example, domestic violence is under-prosecuted and under-punished in Pakistan. The reports described domestic abuse that included the killing and mutilation of women suspected of adultery, as well as other horrific forms of violence, such as ‘oven burns’ that often lead to death, which are inflicted on married women over issues such as inadequate dowries.
Khawar’s application for a protection visa was refused by a delegate of the Minister for Immigration and Multicultural Affairs. Khawar sought review of this decision by the Refugee Review Tribunal, which affirmed the decision of the delegate. The Tribunal did not make findings of fact because it found that even if Khawar’s claims of abuse were true, those harming her were motivated by purely personal considerations. These included the circumstances of the marriage, the fact that Khawar brought no dowry to the family and a dislike of Khawar as an individual. Khawar was not harmed for a Convention reason and, in particular, Khawar was not harmed because she was a member of any particular social group proposed. These included ‘women’, ‘married women in Pakistan’, ‘married women in Pakistan without the protection of a male relative’, ‘married women in Pakistan separated from one’s husband and without the protection of a male relative’, ‘married women in Pakistan suspected of adultery’ or ‘women who have transgressed the mores of Pakistani society’. The Refugee Review Tribunal assumed that the alleged persecution consisted solely of the domestic violence and found that the Refugee Convention was not intended to provide protection to people involved in personal disputes. The attitude of the Pakistani police could not turn harm inflicted by non-state actors for non-Convention reasons into persecution on a Convention ground.
Khawar appealed to the Federal Court of Australia. She argued that the violence inflicted by her husband was condoned or tolerated by the state authorities; that the state authorities failed to provide protection under the law to married women who were the victims of domestic violence; and that such a failure constituted an absence of effective protection by the state by reason of Khawar’s membership of a particular social group, namely married women in Pakistan. The Minister did not challenge that, as a matter of law, women in Pakistan could be a particular social group, but submitted that this was a question of fact for the Refugee Review Tribunal to determine. However, the Minister did challenge that harm by a non-state actor for non-Convention reasons could qualify as Convention-based persecution where the state withholds effective protection on a Convention ground.
The primary judge, Branson J, allowed Khawar’s appeal and sent the case back to the Refugee Review Tribunal for further findings of fact. Justice Branson relied upon the decision of the House of Lords in Islam/Shah, which was delivered after the decision of the Tribunal in Khawar. She found that the systematic refusal by the police to protect members of a particular social group from violence was alone capable of amounting to persecution under the Refugee Convention. The Minister appealed to the Full Court, where a majority dismissed the appeal. Justice Lindgren, with whom Mathews J agreed, held that persecution could be found in a state’s systematic failure to provide protection alone or in the combined effect of the husband’s violence and the state’s systematic failure to provide protection. Justice Hill, in dissent, found that mere inaction by the state could not, by itself, constitute persecution. Further, if the relevant social group were defined as Pakistani women, Khawar did not fear persecution because of her membership of that social group because not all women or married women in Pakistan were potential victims of persecution. Alternatively, if the social group were defined as women whose husbands abused them, the social group was impermissibly defined by reference to the persecution feared.
The Minister appealed to the High Court arguing that the failure of state protection was not capable of amounting to persecution, whether it was considered alone or in combination with the husband’s violent conduct. Again, the Minister did not challenge that, as a matter of law, women in Pakistan could be a particular social group. The High Court’s decision focused on two main questions: whether a lack of state protection could amount to persecution; and whether Khawar was a member of a particular social group.
Khawar argued that her persecution resulted from a lack of state protection against domestic violence. The first counter-argument raised by the Minister was that persecution and lack of protection are distinct concepts in the Convention definition of a refugee that should not be conflated. Article 1A(2) defines a refugee as a person who has a well-founded fear of being persecuted for a Refugee Convention reason and who is outside the country of his nationality and is unable or, owing to such fear, is unwilling to avail himself of the protection of that country. Thus the Minister argued that the Refugee Convention definition of a refugee has two limbs, persecution and lack of state protection, which are cumulative requirements. The existence of or reasons for lack of state protection cannot be equated with the existence of or reasons for persecution. Chief Justice Gleeson and McHugh and Gummow JJ concluded that the reference to ‘protection’ in the second limb of the Refugee Convention definition refers to diplomatic or consular protection extended by a country to its nationals abroad. However, this does not mean that protection provided within a state is not relevant to the notion of persecution under the first limb of the Refugee Convention definition.
The second counter-argument raised by the Minister was that persecution could not include a mere failure to protect because persecution is an active concept, whereas a failure to protect is merely an omission. The Minister relied on the Macquarie Dictionary, which defines persecution as: ‘to pursue with harassing or oppressive treatment; to oppress with injury or punishment for adherence to principles; and to annoy by persistent attentions, importunities, or the like’. Further, the withholding of police protection did not itself give rise to Khawar’s fear of persecution because it would have had no consequences in the absence of the threatened harm from her husband. The Minister contrasted this with Chen Shi Hai v Minister for Immigration and Multicultural Affairs, where the denial of food, shelter and medical care normally provided by the state caused a direct and significant harm to the persons affected. While the denial in Chen amounted to positive action directly harming the claimants, this case involved mere inaction that had no consequences without the violence by the husband.
Four of the five judges of the High Court found that persecution could occur in situations where there was a lack of state protection against domestic violence. However, the majority split 2:2 on whether to analyse this case as persecution consisting of the lack of state protection alone, or as the combination of the husband’s violence and the lack of state protection. Justices McHugh and Gummow analysed the case on the first basis and expressly found it unnecessary to consider the second approach. Chief Justice Gleeson and Kirby J both analysed the case on the second basis. Chief Justice Gleeson did not consider the first approach, whereas Kirby J expressly rejected it. Justice Callinan, in dissent, held that persecution would require at least a deliberate state abstention, rather than mere state inaction. He found there was no persecution on either approach.
Justices McHugh and Gummow found that persecution could be constituted by a lack of state protection alone. In Chan, Mason CJ recognised that persecution by a state may be constituted by the ‘denial of fundamental rights or freedoms otherwise enjoyed by nationals of the country concerned’. Khawar complained she was unable to obtain police protection in respect of domestic violence. This selective or discriminatory treatment amounted to a denial of a fundamental right to protection otherwise enjoyed by Pakistani nationals. Accordingly, ‘the persecution in question lies in the discriminatory inactivity of State authorities in not responding to the violence of non-State actors. Thus, the harm is related to, but not constituted by, the violence.’ The domestic violence was a precondition for, but not a part of, the persecution under the Refugee Convention. They also held that Khawar would need to prove that the police failed to provide protection to her for a Convention reason. It would not be sufficient to show that there had been a systematic failure of law enforcement due, for example, to a shortage of resources.
By contrast, Gleeson CJ found persecution might result from the combined effect of the conduct of two or more agents, such as private individuals and the state. A relevant form of state conduct in these circumstances might be a state tolerating or condoning the infliction of serious harm by a non-state actor in circumstances where the state has a duty to provide protection against such harm. The requirement that the persecution must be for reason of one of the Convention grounds may be satisfied by the motive of either the state or non-state actor. Consequently, even if Khawar’s husband abused her for personal reasons, she may still fear persecution for a Convention reason if the state withheld police protection by reason of her membership of a particular social group.
Justice Kirby also adopted the combination analysis. As to whether persecution requires an action rather than an omission, Kirby J noted that words in international conventions are not susceptible to exposition simply by reference to English and Australian dictionaries. Persecution may occur where there is a risk of serious harm by a non-state actor and the state fails to protect the basic human rights and dignity of its nationals. Unless there is a complete breakdown in the state, decision-makers are entitled to assume that the state is capable of protecting their nationals. Justice Kirby adopted the concise formula ‘Persecution = Serious Harm + Failure of State Protection’. In determining whether the persecution was ‘for reasons of’ a Convention ground, the nexus between the persecution and Convention reason may be provided by either the serious harm limb or the failure of state protection limb. Thus the Refugee Convention reasons for persecution may be satisfied by the motive of either the state or non-state actor.
Justice Callinan held that persecution involves deliberate action rather than mere inaction and that the evidence fell well short of establishing that the state was complicit in violence against women in abusive relationships. The law provides, if enforced, adequate protection in the case of domestic violence. If there is no government policy that police ignore calls for help from victims of domestic violence, then there is no persecution. It would be difficult, or even impossible, for an Australian court to assess the policing priorities of other countries. Justice Callinan also distinguished Chen from the present case because it involved the deliberate abstention by the state from the provision of necessities that were routinely provided to others, rather than a mere reluctance to intervene.
The Refugee Convention was not designed to confer a general right of asylum upon people suffering hardship. Instead, Khawar needed to establish that she had a well-founded fear of persecution for reasons of one of the Convention grounds. The Minister argued that if the social group was broadly defined as ‘women in Pakistan’, then Khawar did not fear persecution by reason of her membership of that group because only women with abusive husbands fear the systematic failure of police protection for victims of domestic violence. On the other hand, if the group was defined narrowly as ‘women in Pakistan with abusive husbands’, then it would be impermissibly defined by the persecution feared.
Chief Justice Gleeson took the boldest approach by concluding that it would have been open to the Refugee Review Tribunal to conclude that ‘women in Pakistan’ were a particular social group. Justices McHugh and Gummow held that it would have been open to the Tribunal to conclude that there was a social group in Pakistan comprising, at its narrowest, married women who lived in a household without a male blood relative to whom the woman could look for protection against violence by members of the household. They did not find it necessary to determine whether any of the wider social groups could be made out, but stated that nothing in their reasons was intended to foreclose such a finding. Similarly, Kirby J found it unnecessary to decide whether women in Pakistan could constitute a social group because he found that a narrower formulation could be made out on the facts. Justice Callinan, in dissent, would not have recognised any of the suggested groups as being particular social groups.
One objection raised to characterising women as a particular social group was the size of such a group, as women make up around half of the population of any state. Chief Justice Gleeson and McHugh, Gummow and Kirby JJ all stated that the large size of a group does not necessarily mean that it is not a particular social group. According to Gleeson CJ, it is power rather than numbers that creates the conditions in which persecution may occur. This means that victims of persecution in a country may be a majority, though the large size of a group might make the suggestion of persecution implausible in some cases. According to McHugh and Gummow JJ, the inclusion of race, religion and nationality in the Refugee Convention definition shows that size of itself can be no objection to the finding of a social group. Justice Kirby stated that the large size of a group would not of itself mean that there was no relevant social group, as evidenced by the persecution of six million Jews under Nazi rule. By contrast, Callinan J found that regarding half of the population of a state as capable of constituting a social group ‘a somewhat unlikely proposition’. Further, a particular social group must be part of something and less than a whole, so it must be a specific and readily definable group or social unit. Justice Callinan held there needed to be at least one characteristic, attribute, activity, belief, interest or goal that unites the members of the group and makes the group identifiable. He doubted whether groups such as ‘married women in Pakistan’, could be identified with sufficient precision.
The Court also considered the need to identify the particular social group independently of the persecution suffered. Following Applicant A v Minister for Immigration and Ethnic Affairs, the particular social group cannot be defined solely by reference to the persecutory conduct, though the persecution may serve to identify or even cause the creation of such a group. Allowing persecutory conduct of itself to define a particular social group would allow ‘particular social group’ to become a catch-all phrase and would impermissibly weaken the cumulative requirements of the definition of a refugee because any group that was persecuted would be persecuted for reasons of their membership of a particular social group. In Applicant A, couples who wanted to have more than one child in China were held not to be a particular social group because they were not linked by any attributes or characteristics apart from being persecuted under the one-child policy. In Khawar, Gleeson CJ held that women in any society are a distinct and recognisable group, which exists independently of the manner in which they are treated, either by males or by governments. Similarly, McHugh and Gummow JJ concluded that the narrower group of married women, living in a household without the protection of a male blood relative, existed independently of the persecution suffered.
The High Court’s decision in Khawar challenges gendered public/private distinctions in international refugee law because it recognises that a state may bear public responsibility for failing to protect a person from private harms. This should be contrasted with earlier decisions, in Australia and other jurisdictions, where domestic and sexual violence have been characterised as private harms perpetrated for personal reasons. For example, in Minister for Immigration and Multicultural Affairs v Ndege, Weinberg J found that the real source of persecution in a case involving domestic violence was the violent husband and not the state inaction in the face of that violence. As the husband inflicted violence on his wife for personal reasons, such as revenge, there was no persecution on a Refugee Convention ground in which the state could be complicit.
A more extreme example is Campos-Guardado v Immigration and Naturalization Service, where the Fifth Circuit Court of Appeals in the United States upheld the denial of asylum to a Salvadorian woman who had been attacked and raped when a paramilitary group stormed her uncle’s house. The armed assailants tied the woman up and forced her to watch as they hacked flesh off the bodies of her male cousins and uncle before shooting them. Two assailants then raped the claimant and her female cousins, while a third shouted political slogans. Later, one of the attackers threatened to kill the claimant if she told anyone about the attack. The claimant argued that she had been persecuted on the grounds of imputed political opinion and membership of a particular social group, namely her family. The Court upheld the Board’s decision that while the attackers may have been motivated by their own political goals in attacking the male relatives, the rapes were of a personal nature.
The approach in Khawar and Islam/Shah provides a way to bridge the gap between harm perpetrated by a private actor for personal reasons and a state being unable or unwilling to protect against such harm for a Refugee Convention reason. This is important for women because many of the harms women suffer occur in the private sphere and states have often been unwilling or unable to protect women from these harms. The majority in Khawar reinforces the view that states have an obligation to refrain from directly harming their citizens and a duty to protect citizens from violations of their rights by non-state actors. Persecution can thus occur at the hands of state and non-state actors, and state involvement in this persecution can take the form of state action or inaction.
A similar notion of state responsibility exists in international human rights law. For example, the International Covenant on Civil and Political Rights requires states to ‘respect and ensure’ fundamental rights provided in the Covenant. In the Velásquez-Rodriguez Case (Honduras), the Inter-American Court of Human Rights found that this duty extends to states having an obligation to protect against human rights violations and to investigate, prosecute and punish non-state actors who have violated human rights. This obligation goes beyond states ‘respecting’ rights; it requires states to ‘ensure’ rights are respected by non-state actors. A state may fail in its duty to ensure the rights of its citizens if human rights violations occur with its support or acquiescence, or if the violations occurred because of the absence of adequate preventative or punitive measures. Likewise, the Declaration on the Elimination of Violence against Women calls on states to pursue, by all appropriate means and without delay, a policy of eliminating violence against women. This includes exercising due diligence to prevent, investigate and punish acts of violence committed against women, whether committed by state or non-state actors.
Refugee protection will not, however, be granted every time an individual suffers harm or discrimination by a non-state actor. While some feminist scholars challenge public/private distinctions, they do not necessarily seek to collapse all public/private distinctions in order to make everything a matter for legal regulation. For example, Mathew argues that a couple’s choice about the number of children they want should fall within the private sphere rather than being subject to governmental control through policies such as China’s one-child policy. The problem with the public/private distinctions is not that there should be no distinction between the public and the private, but rather that the line between the two is often drawn in an arbitrary way that excludes women’s concerns from the public sphere. Instead of ridding the law of all public/private distinctions, Charlesworth and Chinkin argue that it is important to redraw the lines between the public and private spheres in a way that does not systematically marginalise or exclude women’s experiences. However, it can be difficult to determine where this line should be redrawn in practice.
In this respect, it is important to note that the decision in Khawar does not mean that anyone who suffers harm in the private sphere will be granted refugee status. Instead, the generality of the decision will be checked in its application to future cases on at least three levels. First, the harm or discrimination in question must reach a sufficient level of seriousness to constitute ‘persecution’. Not all forms of gender-based discrimination or gender-specific harms are sufficiently serious to constitute persecution. For example, a woman may receive unwanted attention because of her sex, such as being whistled at in the street, but this is not sufficiently serious to justify refugee status. The Refugee Review Tribunal in Khawar assumed for the sake of argument that the domestic violence in question was sufficiently serious to amount to persecution, but this must still be proved on the facts. Even if the domestic violence and lack of protection in Khawar were held to be sufficient to constitute persecution, this does not necessarily mean that all domestic violence would amount to persecution. The High Court has also not yet addressed whether other gender-specific forms of harm, such as rape and female genital mutilation, would be sufficiently serious to amount to persecution.
Justices McHugh and Gummow noted that while persecution involves discrimination, not all discrimination amounts to persecution. Chief Justice Mason stated in Chan that the denial of fundamental rights or freedoms enjoyed by other nationals may constitute persecution, but other forms of selective or discriminatory treatment may not be sufficient to amount to persecution. According to the majority in Chen, whether differential treatment amounts to persecution depends on whether it offends the standards of civil societies. In Khawar, the denial of law enforcement to protect against violence to the person represented a denial of a fundamental right otherwise enjoyed by Pakistani nationals. However, the High Court did not consider whether discrimination against women in other contexts, such as dress codes and travel restrictions, would amount to persecution, either considered alone or cumulatively.
Second, the obligation on states to provide protection against harm by non-state actors in refugee law is not an absolute obligation to provide protection against harm in all circumstances. All victims of crime experience a failure of state protection but not all victims of crime would be granted refugee status. Persecution is concerned with sustained or systematic violations of basic human rights. Consequently, a state must undertake its duty to protect its nationals from harm with ‘due diligence’, which means that it must undertake its duties seriously. If a state has laws against domestic violence and has taken reasonable steps to enforce those laws without discrimination, then it is unlikely that an applicant will be granted asylum. According to Lord Steyn in Islam/Shah, domestic abuse of women is prevalent in many countries and does not, by itself, justify refugee status. The distinctive feature in that case was that women in Pakistan were unprotected by the state. Thus, while violence against women occurs in all countries, the real issue is whether the state makes a genuine effort to provide effective protection against, and remedies for, such violence.
However, the decision in Khawar provides little guidance about what constitutes adequate steps to protect women against violence. One paradox that arises is how women who fear domestic violence in one state can be granted refugee status in another state when domestic violence exists everywhere and no state offers complete protection from it. Refugee law is sometimes characterised by a distinction between developing states that produce refugees and developed states that receive them. This implicit ‘us/them’ dichotomy may make it easier to establish gender persecution in cases involving non-western practices, such as female genital mutilation. Domestic violence challenges the self-perception of western states as non-refugee producers because it is prevalent in western states and their own protection mechanisms may be inadequate. However, while no state offers complete protection from domestic violence, there may still be significant differences in the way various states respond to the problem. Granting asylum may still allow a claimant to seek out better protection than that from which she currently ‘benefits’. An abused wife may also have a significantly better chance of escaping persecution if she can escape from her violent husband, even if she seeks asylum in another state that does not have a perfect record for protecting women.
It is difficult to determine whether a state has taken adequate steps to provide protection, particularly where that protection has ultimately been unsuccessful. Simply importing the due diligence test from human rights law into refugee law may be problematic. International human rights law is concerned with whether a state should be held accountable or responsible for violations of human rights standards. By contrast, refugee law is concerned with whether the international community should offer surrogate protection because there has been a failure of state protection. Should a woman who has suffered domestic violence be denied refugee status where a state attempted to provide protection but was ultimately unsuccessful? Should her refugee status depend on whether the state’s attempt at protecting her was undertaken in good faith? In Canada, the case law supports the proposition that state protection must be adequate although not necessarily perfect, but there is no consensus on whether ‘adequate’ means effective protection or might be satisfied by a lower standard. Even when a state has taken reasonable measures to protect women as a group from domestic violence, there may still be a case for surrogate protection under refugee law for particular women to whom the state has been unable to provide effective protection. Thus, there may be persecution under refugee law where a state fails to provide effective protection even if the state cannot be held accountable for that failure under international human rights law.
The conventional understanding of the obligation on states to afford protection is that a state may not have taken sufficient steps to protect its nationals if it enacts persecutory laws or if it applies general laws in a persecutory fashion. Thus, it is important to examine both the law in theory and how it is applied in practice. Laws may be inherently persecutory if they result in a denial of basic rights or protection to people of a particular race, ethnicity or gender. For example, Pakistan’s Hudood laws place unjustifiable burdens on a woman claiming rape, as she must corroborate her complaint with the testimony of four male witnesses and failure to prove rape leaves her vulnerable to prosecution for adultery or fornication. The absence of laws may also result in persecution. For example, in In re Kasinga, a young woman was granted asylum because the state had enacted no legislation against female genital mutilation. Laws may also be applied in a persecutory manner, such as only being applied against perpetrators of a particular race, ethnicity or gender, or not being applied to help victims of a particular race, ethnicity or gender. For example, in In re O-Z- & I-Z, the Board found that while the Ukrainian government publicly condemned anti-Semitism, in practice it failed to prosecute breaches of its anti-Semitic laws.
The Canadian Gender Guidelines state that a woman cannot claim refugee status solely on the basis that she is subject to a national law or unofficially sanctioned policy to which she objects. Instead, she would need to show one of the following. First, the policy or law is inherently persecutory. For example, marital rape exemptions in the criminal code may be inherently persecutory. Second, the policy or law is used as a means of persecution for one of the enumerated reasons. For example, prosecution for an offence against the public order for distributing pamphlets might be used as a vehicle to persecute an individual based on the political content of the publication. Third, the policy or law, although having legitimate goals, is administered through persecutory means. For example, it is arguable that the People’s Republic of China is enforcing a legitimate goal of population control through persecutory means such as forced abortions and sterilisations. Finally, the penalty for non-compliance with the policy or law is disproportionately severe. For example, women in Iran are required to wear a chador and non-compliance is a criminal offence punishable by 75 lashes.
In Applicant A, McHugh J held that ordinarily the enforcement of a generally applicable criminal law will not constitute persecution of a social group constituted by those against whom that law is enforced. However, in Chen, Gleeson CJ and Gaudron, Gummow and Hayne JJ noted that ‘[l]aws or policies which target or apply only to a particular section of the population are not properly described as laws or polices of general application’. For example, laws that target or impact adversely upon a particular class or group, such as ‘black children’, as distinct from children in general, cannot be properly described as laws of general application. In Islam/Shah, Lord Millett, in dissent, found that the applicants were not persecuted because they were women, but instead were persecuted because they had transgressed social norms and there was no evidence that men who transgressed different social norms were treated more favourably. However, laws may be inherently discriminatory if they place unjustified restrictions, or disproportionately severe penalties, on one sex.
Due to the factual basis of the case, the High Court did not consider whether state protection must be sought in order to establish that the state was unwilling or unable to offer protection. In many states, women may believe that it would have been futile to seek the protection of the state because the state generally refuses to protect victims of domestic violence. Women who are victims of domestic violence may also risk increased violence by their husbands if they seek protection from the state. In Ward, La Forest J stated ‘it would seem to defeat the purpose of international protection if a claimant would be required to risk his or her life seeking ineffective protection of a State, merely to demonstrate that ineffectiveness.’
Finally, the asylum seeker must establish that the persecution feared was for a Refugee Convention reason. Where the harm perpetrated was by a non-state actor, the majority held that it would not be enough to show that the state failed to protect the asylum seeker because the police were corrupt, inefficient, lazy or under-resourced. This would be insufficient because the lack of protection would not be for a Refugee Convention reason. Instead, following the reasoning of Chief Justice Gleeson and Justice Kirby, it must be shown that either the non-state actor violated the basic rights of the asylum seeker for a Convention reason or that the state failed to protect the asylum seeker for a Refugee Convention reason. Following Justices McHugh and Gummow’s analysis, it must be shown that the state failed to intervene for a Convention reason. Both the systemic nature of the persecution and the motivation for persecution are questions of fact that may be difficult to prove. In Khawar, the High Court did not decide whether Pakistan systematically discriminated against women through selective law enforcement, but instead left that as a question of fact for the Refugee Review Tribunal to determine.
While Khawar helps to bridge some gendered public/private divides, other problems in the presentation and analysis of gender claims in refugee law remain unresolved. This section outlines some areas where there is scope for further development of a gender analysis of refugee law in future refugee cases.
On one level, the decision in Khawar reinforces gendered public/private distinctions because it is premised on the view that harm in the private sphere is perpetrated for personal reasons. The case was conducted on the basis that the domestic violence Khawar suffered at the hands of her husband and his family was motivated by purely personal considerations, including that Khawar brought no dowry to the marriage and the family disliked Khawar as an individual. Khawar attempted to prove her refugee status by arguing that the state did not protect her from domestic violence because she was a woman, not by arguing that she was subjected to domestic violence because she was a woman. Domestic violence is a gender-specific harm because it more commonly befalls women. Yet courts do not generally analyse domestic violence as gender-related persecution, that is violence perpetrated because the victim was a woman. However, the assumption that domestic violence occurs for purely personal reasons is problematic.
In determining whether a gender-specific harm also amounts to gender-related persecution, one must examine the role that the victim’s gender played in the persecution. In some cases, the persecutor would have persecuted the victim regardless of the victim’s gender, even though the method of persecution chosen may be related to the victim’s gender. For example, a person might persecute a couple for political reasons by torturing the man and raping the woman. This may amount to persecution for a Refugee Convention reason, such as political opinion. However, the woman would probably not have been persecuted for gender-related reasons even though she was harmed in a gender-specific manner. However, if the persecutor would not have persecuted the victim had the victim not been of that gender, then an inference may be drawn that one of the motivations for persecution was the victim’s gender. For example, if a person only attacks women, then one of the reasons that a given woman is attacked is because she is a woman. Gender-specific harm and gender-related persecution are not necessarily related, but there is often a link between the two.
While domestic violence is generally perpetrated by men against women, it is usually assumed to occur for personal reasons. The first argument in favour of this approach focuses on the apparently personal triggers for any particular act of domestic violence. For example, domestic violence might be triggered by a wife being disobedient or having an inadequate dowry. However, this reasoning is problematic because many of these triggers for domestic violence reflect assumptions about the appropriate role and behaviour of women rather than being motivated by purely personal considerations. If a husband beats his wife because he is upset that she has been disobedient or that she brought an inadequate dowry to the marriage, then this reflects gendered expectations about the role and status of women in society rather than being purely personal considerations.
Even when the triggering events are not based on gendered considerations, focusing only on the triggers for domestic violence does not adequately explain the underlying reasons for domestic violence. A triggering event may explain why a husband gets annoyed or angry at any given moment, but it does not explain why he believes that beating his wife is an acceptable response. Domestic violence has a gender-specific outcome because it is a manifestation of socially ingrained beliefs about the correctness of male domination and female subordination. While the trigger for any act of domestic violence may or may not be personal, the broader motivation behind domestic violence often lies in the deeply held belief that women are inferior and that men have the right to treat them as they wish. According to Macklin, ‘the issue is not the proximate reason for the violence, but the underlying assumption that men are entitled to beat women’. In many cases, this belief may be so imbedded that the perpetrator does not even consciously register this belief as motivating his actions. Thus, both the triggering event and the husband’s reaction that led to domestic violence, may reflect gender-related rather than purely personal considerations.
The second reason why domestic violence is often discounted as being motivated by personal considerations is that, while domestic violence is generally perpetrated by men against women, many men beat their wives without beating women in general. Thus it is arguable that a husband does not attack his wife because of her membership of a particular social group. However, a man may beat his wife because she is a woman without beating women generally because any act of persecution involves a complex relationship between motivation and opportunity. Lord Hoffman gave a good example of this in Islam/Shah by considering a business rival in Nazi Germany who attacks a Jewish shopkeeper when the state has a policy of not giving protection to Jewish people. On one view, the Jewish shopkeeper was not attacked because of race but because the competitor was motivated by wanting to drive him out of business. However, the competitor would not have acted as he did unless he knew that the authorities would allow him to act with impunity. Thus, one of the reasons the Jewish shopkeeper was attacked was because of his race. In the same way, one reason that a man beats his wife without beating women in general is that he knows that the state will not intervene to help women in domestic disputes. Indeed, many abusive men are likely to abuse any woman they marry, without the motivation for the violence being particular to the woman.
The intimate connection between motivation and opportunity in persecution was not appreciated in Khawar. During the course of argument, the Minister submitted that Khawar had been harmed by her husband for purely personal reasons, such as bringing an insufficient dowry to the marriage, and not because she was a member of a particular social group of women in Pakistan. Chief Justice Gleeson countered with the observation that ‘those males are certainly behaving towards her in a manner to which they would not behave towards another male’. Counsel for the Minister agreed, but stated that the motivation was not connected to any Convention group because ‘[t]he persecutor in the present case does not persecute women generally, [and] does not commit acts of violence against women generally. The persecutor attacks his wife … for personal reasons unconnected to membership of a group.’
However, a racist person may attack individual members of a race, on the basis of their race, without attacking all members of that race. The individuals chosen as victims are likely to be people whom the persecutor comes into close contact with, or who are particularly vulnerable because of their relationship with the persecutor, or who are unlikely to be protected by the law. Examples of this might be found in the treatment of Afro-American slaves where there was an intersection of racially motivated persecution with opportunity because the victims were in a vulnerable position and were unlikely to be protected by the law. In the same way, a violent husband may attack his wife without attacking women generally because of the close proximity of their relationship and the vulnerability of a wife whom the law is unlikely to protect. Domestic violence represents the concurrence of motivation to harm women because they are women, with the opportunity to harm one’s wife because the state does not intervene to protect women in the domestic sphere.
Identifying the reasons for persecution is a difficult task not only because it requires an analysis of motives, but also because persecution may be motivated by two or more considerations. The intersection of gender and other motivations for persecution occurs in other areas of the law, such as rape under international humanitarian law. In Kunarac, the International Criminal Tribunal for the former Yugoslavia (ICTY) held that the instances of rape in that case were committed against Muslim women ‘on the basis only of their Muslim ethnicity’ and that Muslim men and women were killed, raped or severely beaten and the ‘sole reason for this treatment … was their Muslim ethnicity’. However, if the only motivation for raping victims were Muslim ethnicity, then one would expect that Serbian men would have raped Muslim men and women without discrimination, which they did not. Thus, focusing only on the ethnic-specific nature of the rapes ignores their gender-specific nature. Yet if the victim’s gender were the only motivation for persecution in Bosnia, then one would have expected Serbian men to have systematically raped Bosnian Muslim women and Bosnian Serbian women, which they did not. Thus, focusing only on a victim’s gender obscures the ethnic-specific nature of the rapes.
There may be more than one motivation for action and Serbian men may have raped Bosnian Muslim women because they were Muslim and because they were women and because they were living in Bosnia. Likewise, women may be raped as a form of genocide in order to propagate the perpetrator’s racial or ethnic group and to destroy the victim’s racial or ethnic group. In these cases, women are targeted for persecution on the basis of their race and also because of their reproductive capacity as women. To focus only on race or ethnicity or only on gender would obscure who is doing what to whom and why. The law struggles to recognise multiple motivations for harm and gendered motivations behind crimes are often ignored when other motivations can also be identified. Crenshaw argues that in order to accurately capture the nature of harms perpetrated against women of a particular race or ethnic minority, the law needs to recognise the intersectional nature of these women’s identities. Women may experience distinct vulnerabilities created by their numerous overlapping identities. Green argues that experiences of women in war are often shaped by their identity as women and as members of a particular national, ethnic, racial or religious group. For example, hate propaganda was used in Rwanda before the genocide to incite sexual violence against Tutsi women.
There has been a growing awareness that women may be persecuted in wars because they are women. For example, in Delalic, the Trial Chamber of the ICTY found that targeting women for rape because they were women constituted a discriminatory purpose for the offence of torture. The definition of crimes against humanity in the Rome Statute for the International Criminal Court also includes gender as a basis for persecution against an identifiable group or collective. However, the claim that women are attacked because they are women is still highly contentious. For example, Patel claims that women in Bosnia were raped because they were Bosnian and because they were women, so both bases should be recognised in determining their refugee status. However, Spijkerboer argues that rape in war usually does not amount to gender-related persecution because gender only affects the means of persecution used rather than causing the persecution. If a woman who was vaginally raped would have been persecuted in another way had she been a man, then the crime may be gender-specific but it would not amount to gender-related persecution. However, if that woman would not have been subjected to persecution had she been a man, then the persecution may be gender-related. It is also arguable that the sexual nature of the violence perpetrated against women represents an expression of hatred against women. After all, if the motivation for harm were gender-neutral, why would the harm take a gender-specific form? These questions of motivation and mixed motivations are extremely difficult to resolve and require further attention in refugee law and international humanitarian law.
The relationship between gender-specific harms and gender-related persecution is also often ignored because of the complex role that gender plays in such cases. On the one hand, women asylum seekers must demonstrate that they have suffered ‘serious harm’. Gender-specific harms have often been trivialised and relegated to the private sphere. Feminist scholars have sought to address this problem by attempting to de-emphasise the importance of gender, for example, by arguing that violence in the home is not conceptually different to violence in the public sphere. On the other hand, if asylum seekers want to argue that these harms amount to gender-related persecution, they need to re-emphasise the importance of gender, for example, by arguing that domestic violence predominantly involves violence by men against women because they are women. While there is no conceptual problem with this approach, it causes strategic problems in practice because the importance of gender is de-emphasised at the first hurdle of refugee claims and then re-emphasised at the final hurdle. The same issue arises within feminism more generally. For example, MacKinnon argues that rape is a crime of violence rather than an issue about sex and that this characterisation is important in order that rape not be dismissed as simply a personal issue. However, using a neutral, abstract word such as violence removes the fundamental role sex plays in the violence, which obscures the gendered nature of the crime.
This raises the problem of pursuing gender-related persecution under the gender-neutral category of ‘particular social group’. Some commentators argue that gender should be added as a sixth category under the refugee definition because of the importance of naming gender-related persecution as a gender issue. Other commentators, most notably Macklin, argue that it is preferable to pursue gender-related persecution under the existing categories for a variety of reasons, including that one should not collapse every instance of persecution of women into the category of persecution on the grounds of gender. The conflict between these views reflects a tension more generally within feminism about the merits of incorporating gender claims into existing legal categories. For example, in international humanitarian law, there is debate over whether rape should be charged as a form of genocide and torture. On the one hand, this approach has allowed rape to be recognised and prosecuted as a weapon of war, rather than being viewed as just an unfortunate and unavoidable by-product of war. On the other hand, prosecuting rape as torture or genocide obscures the gender-specific nature of rape because it is prosecuted as a gender-neutral crime. This is problematic because rape tends to be committed by men against women and the sexual nature of rape is integral to the experience of victims and to the stigma associated with the crime. Feminist scholars often face this tension between emphasising the gendered nature of a crime and risking that crime being marginalised as a ‘women’s issue’, or attempting to mainstream the issue by emphasising the similarity between that crime and existing gender-neutral crimes and risking its gender-specific nature being obscured.
There are three important reasons why we should reject the assumption that harms perpetrated in the private sphere are perpetrated for personal reasons. First, far from being purely personal, violence in the private sphere is profoundly political. According to Bunch, such violence ‘results from the structural relationships of power, domination, and privilege between men and women in society. Violence against women is central to maintaining those political relations at home, at work, and in all public spheres.’ The political nature of violence against women is yet to be fully recognised by the law. For example, in Refugee Appeal No 71427/99, the New Zealand Refugee Status Appeals Authority cursorily dismissed the possibility that the husband persecuted his wife because she was a woman on the basis that such a suggestion was artificial. Despite recognising that domestic violence against women is a manifestation of historically unequal power relations between men and women, the Authority simply asserted that this did not explain why that particular claimant was at risk of serious harm. However, in Lazo-Majano, the Ninth Circuit Court of Appeals found that a Salvadorian army sergeant who raped and beat his domestic house worker was asserting a political opinion that men have a right to dominate women and that his statements and actions demonstrated a generalised animosity to the opposite sex.
Second, non-state actors should be able to provide the Refugee Convention reason for persecution. Following the approach of Gleeson CJ and Kirby J, where persecution results from the combined conduct of persecution by a non-state actor and lack of protection by the state, the Convention reason for the persecution may be provided by either the state or the non-state actor. Thus, courts should not simply inquire whether the state withheld protection to women suffering domestic violence because they were women. Courts should also inquire whether the violent husband persecuted his wife because she was a woman or for some other Convention reason. This line of inquiry is not available under the approach taken by McHugh and Gummow JJ because they limit the concept of persecution to the lack of state protection alone, so it is only relevant to consider whether the lack of protection is motivated by a Convention reason.
Finally, the formulation that persecution can occur when a state is unable or unwilling to protect against harm by a non-state actor only makes sense if the non-state actor may provide the Convention reason for persecution. A state may be unwilling to provide protection due to a Convention reason, for example not providing protection to Jewish people because they are Jewish. However, it is hard to imagine a situation where a state is unable to provide protection due to a Convention reason. Inability to provide protection typically refers to situations where there has been a breakdown in a governmental authority so that the state is unable effectively to police crime. Inability in this sense is not motivated by Convention reasons. A state may claim that it is unable to help certain victims because it lacks resources, but this often reflects the fact that those victims are not given priority because of their race or gender, so the state is actually unwilling rather than unable to provide protection. However, if a husband persecutes his wife for a Convention reason, then there will be persecution provided the state is unwilling or unable to protect her. It would not matter whether the state’s inaction is motivated by a non-Convention reason, such as the lack of sufficient resources to protect victims.
The High Court did not resolve whether women in Pakistan could constitute a particular social group. The Minister accepted that, as a matter of law, women in Pakistan could constitute a particular social group. However, only Gleeson CJ held that it would be open to the Refugee Review Tribunal to conclude that women in Pakistan were a particular social group. Justices McHugh and Gummow, along with Kirby J, did not find it necessary to determine the issue because they found that the applicant could have been a member of a much narrower social group, such as married women who lived in a household without a male blood relative who could protect her from violence by other household members. Justice Callinan held that neither women in Pakistan nor any of the subsets suggested could amount to particular social groups. Thus the decision in Khawar is not as sweeping as the decision in Islam/Shah in the United Kingdom where three of the five Lords accepted that women in Pakistan constituted a particular social group, or the decisions in Ward in Canada and In re Acosta in the United States, which both listed ‘sex’ as an example of a particular social group. While the majority of the High Court did not foreclose the possibility that women in Pakistan could be a social group, they did not affirmatively decide the issue.
The best formulation of the relevant social group is simply ‘women’. The reference to a particular state is unnecessary because women form a particular social group in any state, though women are not necessarily persecuted in all societies because they are women. By analogy, people who follow a particular religion, such as Judaism, form a particular social group in any society but they are not necessarily persecuted in all societies because of their religion. Women are a clearly identifiable group with distinctive attributes and characteristics. Women do not need to be a cohesive group in order to be a particular social group. Chief Justice Gleeson noted that cohesiveness might assist in defining a group but it is not an essential attribute; indeed, some social groups are notoriously lacking in cohesiveness. Further, the group of women is not defined solely by reference to the persecution feared but rather exists independently of the alleged persecution by abusive husbands and the state.
The group is sometimes more narrowly defined as women who are or have been victims of domestic violence in a particular country. This approach incorporates a requirement of past or present persecution into the formulation of the social group, when it is well-established that someone may be a refugee provided they have a well-founded fear of persecution, even if they have not suffered persecution to date. This group may not exist independently of the persecution feared, depending on how the persecution is defined. Following the approach of Gleeson CJ and Kirby J, if persecution consists of the combined effect of the husband’s violence and the state’s failure to protect, then domestic violence should not be used as the only factor to define the group because it is part of the persecution. However, following the approach of McHugh and Gummow JJ, if the persecution is defined as the failure of state protection alone, then defining the particular social group as women who are victims of domestic violence is less problematic because the domestic violence is not part of the persecution feared and thus the social group is not being defined by the persecution. This may be one advantage of the approach of McHugh and Gummow JJ.
Various other subsets of women have been identified as forming particular social groups. For example, McHugh, Gummow and Kirby JJ sought to define the social group narrowly as, for example, married women who live in a household without a male blood relative to whom the woman could look for protection against violence by members of the household. This group probably exists independently of the persecution feared because it is identified by characteristics that mark the group out for persecution rather than the persecution itself. However, it is difficult to see that this is a recognisable group or that the state withholds protection from this group alone, rather than withholding protection from women in domestic situations generally. Justices McHugh, Gummow and Kirby also did not explain why they focused on these factors in order to circumscribe the group of women. Similar approaches have been criticised for simply stringing together a list of all of the qualities, circumstances and features of the claimant that have been factors in leading to her persecution. The Executive Committee of the UNHCR has recognised that women who face harsh or inhuman treatment, due to having transgressed the social mores of their society, may be considered as a particular social group. However, this should not be taken to exclude the possibility that women per se may form a particular social group.
One concern commonly raised about defining the social group broadly as ‘women’ or even ‘women in Pakistan’, is that not all women in Pakistan suffer domestic violence, so it is difficult to see that victims of domestic violence are persecuted because they are women. However, if a social group is not defined by the persecution, then it is almost inevitable that not all members of the group will be persecuted. For example, even though Jews were generally persecuted in Nazi-occupied Europe, some Jewish people were able to escape persecution due to their wealth or other circumstances. The questions of whether a given person is persecuted for reason of their membership of a particular social group, and whether every member of a particular social group is persecuted, are distinct questions that should not be confused.
Justice Hill in the Federal Court argued that ‘women in Pakistan’ could not be a particular social group because although not all members of the group need to be persecuted, all members must be subject to the potential of persecution. He found that while all homosexuals are subject to the potential of persecution, not all women are at risk of persecution because not all women have abusive husbands. Whether a homosexual or a woman is in actual risk of persecution depends on whether the homosexual is ‘out’ and comes into contact with homophobic people or whether the woman marries an abusive man. But members of both groups have the potential to suffer persecution. Justice Hill’s approach is flawed because he defined the persecution purely by reference to the husband’s violence, rather than considering the failure by the police to afford protection as a form of persecution. All women have the potential to suffer persecution because protection will not be given to any woman who suffers domestic violence. This approach is similar to the High Court’s decision in Chen, where it was found that the state withdrew services, such as medical treatment, from a particular social group identified as ‘black children’ in China. The group was not limited to black children who suffered serious illnesses, even though only those children were likely to suffer serious physical harm as a result of the policy.
A second and related concern about recognising women as a social group is that this will open the floodgates and allow half the population of certain countries to gain refugee status. This argument is flawed both in theory and in practice. Theoretically, this argument confuses the fact that every member of the group must have the potential to suffer persecution with the requirement that a particular person claiming refugee status must have a well-founded fear of persecution. While all women have the potential to suffer persecution in the form of lack of state protection in the face of domestic violence, only women who have been or are likely to be victims of domestic violence will have a well-founded fear of that persecution eventuating. By analogy, finding that a particular claimant has a well-founded fear of persecution for reasons of her religion (Judaism) does not mean that every Jewish person has a well-founded fear of persecution and will be granted refugee status.
On a practical level, even though women represent a large number of refugees worldwide, they represent a smaller proportion of refugees resettled overseas. This results from many factors, including women’s lack of resources, lack of mobility, lower earning capacity and inability to travel unaccompanied or without permission. According to the research unit of the Refugee Review Tribunal, domestic violence was an issue in only 76 refugee cases between July 1993 and 31 December 1996. This represents 0.78 per cent of the 9797 decisions made by the Refugee Review Tribunal over that period. The claimants were successful in only 19 of these appeals and Crock concludes that few of these cases could be described as ‘pure’ domestic violence cases, as most involved some element of state action. It is also unlikely that there will be a flood of women refugees following the decision in Khawar. Following the introduction of Gender Guidelines and the decision in Ward in Canada, the total numbers of positive claims decided in accordance with the guidelines were: 78 in 1993, 204 in 1994, 212 in 1995, 150 in 1996 and 104 in 1997. Thus the fear of opening a floodgate would appear to be unfounded.
It is also worth noting that the High Court has not yet considered whether gender-specific harms may also amount to persecution for other Refugee Convention reasons, such as religion and political opinion. While female resistance to oppression and violence is often viewed as a private act committed for personal reasons, it would be open to conclude that a woman who objects to sex discrimination or male domination may fear persecution for reasons of religion or political opinion. According to Spijkerboer, ‘women who fear persecution because they transgress social mores in general are not persecuted because they are women. They are persecuted because they refuse to be “proper” women. When seen in this context one cannot but consider such an act of defiance as being both political and religious’.
Political opinion should not be narrowly construed as simply opinions expressed in the public arena (such as political speeches) or organised forms of public protest (such as marches and street demonstrations). Women and men may both be persecuted for being visibly active in politics or political activities. However, as women have traditionally not taken part in the public sphere, they may also express their political opinions in less-conventional ways, such as rejecting social customs about how women ought to dress and behave. Political activities undertaken by women have also often been mischaracterised as non-political or being political activities of an insufficiently high level, such as hiding people and passing messages. No activity is inherently political, so even apparently non-political activities such as cooking food may be political in particular contexts. For example, setting up communal kitchens under the fascist Pinochet regime could be seen as a form of political resistance. ‘Political opinion’ should not be conceptualised in such a way as to exclude these forms of expression and protest.
In Fatin v Immigration and Naturalisation Service, the Third Circuit Court of Appeals recognised that feminism could qualify as a political opinion under the Refugee Convention. Opposition to laws or social practices that discriminate against women may demonstrate a political opinion about freedom of choice, freedom of expression and the importance of equality for women. Similarly, conflicts concerning the demarcation of privacy, such as whether women should be free to wear the veil or not, free to have an education or undertake certain work, or free to reject male domination, violence or female genital mutilation, are conflicts of a political nature. However, in Fatin, the Court found that a woman would only be persecuted on account of her feminism if her views were so profound that she would choose to suffer the severe consequences of non-compliance rather than conforming to the gendered laws. Other decision-makers have similarly concluded that dress codes do not amount to persecution because women could avoid harsh penalties simply by complying with these requirements, even if that would offend their beliefs. Yet would it be reasonable to conclude that a law requiring Orthodox Jews or Muslims to eat pork would not amount to persecution because these people could easily avoid penalties by eating pork, even if that offended their religious beliefs? A law may be persecutory because it imposes a disproportionate penalty or because it is inherently persecutory by imposing unreasonable requirements on one group.
It is also flawed to narrow one’s attention solely to dress codes. Requiring women to wear the veil is usually only one strand in a web of discriminatory laws and practices that restrict the rights of women. The UNHCR Handbook provides that serious restrictions on the right to earn a livelihood, to practise religion or to access normally available educational facilities may amount to persecution, particularly where such restrictions are cumulative. In many Islamic societies, the veil has become encoded with political meaning about how a government approaches the ‘woman question’ and as a symbol of how it approaches state power, popular morality and religious values more generally. While women should be free to wear the veil, in societies where the veil is compulsory, a refusal to wear the veil, or showing strands of hair or makeup on the face, may represent political statements about the freedom and equality of women. In Fisher v Immigration and Naturalisation Service, the Ninth Circuit Court of Appeals held that the imposition of dress codes on women may amount to sex discrimination but falls short of persecution. But would the same decision be made if Jews were made to wear a yellow Star of David or homosexuals were made to wear a pink triangle? Wearing a particular dress can mark individuals out for persecution and symbolise the general oppression to which they are subject. According to Mathew, the veil may be made compulsory rather than optional in order to suppress dissidence and to clearly identify persons who are dissidents.
As domestic violence is often an expression of a belief about male superiority and female subordination, it is arguable that a woman who resists such violence may be expressing a political opinion rejecting this view. For example, in Lazo-Majano, the Ninth Circuit Court of Appeals found that a Salvadorian army sergeant who raped and beat his domestic house worker was asserting a political opinion that men have a right to dominate women. The Court held that by fleeing the country, the applicant was expressing her political opinion to the contrary. However, this reasoning is problematic because if the applicant only demonstrated her contrary political opinions by escaping, then she was not persecuted for reasons of her political opinion before she escaped. Alternatively, if anyone who objects to being beaten demonstrates a political opinion then anyone who is beaten would be eligible for refugee status. According to Macklin, it is more plausible that men beat women because of what men believe about women, rather than because of what women believe or are imputed to believe about themselves. However, this is not to say that women cannot be raped because of their political opinions. For example, in Campos-Guardado, the asylum seeker was raped while her attackers chanted political slogans. Contrary to the actual decision of the Court, it may have been reasonable to conclude that the attackers were politically motivated and that they believed their victim had contrary political views, for which they were punishing her.
If the suppression of women is justified by a religion, then a woman who opposes such suppression may be persecuted for reasons of religion. Religion as a Refugee Convention ground includes the freedom to hold a particular belief system as well as freedom not to hold a particular belief system. It would also include freedom to practise a particular religion in a particular way and freedom not to practise the prescribed religion in the prescribed way. For example, a woman who refuses to wear the veil in an Islamic society may do so because she follows a different interpretation of the Koran or because she does not follow the Koran at all. If she is persecuted because of this behaviour, then she is persecuted on the grounds of her political or religious beliefs. The Australian Gender Guidelines recognise that where a religion prescribes that women follow certain roles, or codes of dress and conduct, women who do not fulfil those requirements may have a well-founded fear of persecution for reasons of religion. For example, in some countries a woman who refuses to return to an abusive husband may be accused of violating a religious requirement that women be subservient to the will of their husbands.
This raises the contentious relationship between gender persecution, culture and cultural relativism. Culture and religion are often used to justify the denial of basic rights to women, including the right to travel, own property and gain employment, as well as to justify gendered practices, such as dress codes and female genital mutilation. Cultural relativists argue that these differences simply reflect different, but equally valid, cultural attitudes. Following this approach, human rights law involves the unacceptable imposition of western standards, or western standards cloaked as international standards, upon non-western states. As refugee law largely consists of western states determining refugee claims from non-western states, refugee decisions are similarly criticised for being arrogant, hypocritical and for relying on unfair cultural stereotypes. However, in Khawar, Gleeson CJ stated that if a claimant could show state tolerance or condonation of domestic violence, then it would not be an answer to her case to say that such a state of affairs resulted from entrenched cultural attitudes.
International instruments have rejected the claim that discrimination against women can be justified by reference to cultural or religious beliefs. In many cases, the persecution being considered is actually against the law of the state in question. For example, in Khawar, Pakistan had laws outlawing domestic violence and the persecution arose from the police failing to enforce these laws. It would seem disingenuous for a state to claim that enforcing its own laws amounts to cultural imperialism. Even if there was a cultural and legal consensus in a country favouring a particular religious or cultural practice, such as female genital mutilation, refugee law seeks to protect an individual who wishes to separate herself from that consensus. In this way, refugee law does not involve the imposition of western or international views on an unwilling person from a non-western state, even though the decision may be offensive to other people in that state. These women may also be unable to effectively renegotiate cultural or religious norms within their society because of the discrimination they face.
A final problem with persecution on the basis of religion is that religious views are often enforced by non-state actors, such as individuals and religious authorities, rather than by states. However, there may still be a nexus between the persecution and the state if the state fails to protect women from violations of their fundamental rights. Thus, in In re Kasinga, the Court found that female genital mutilation that was not performed by the state still amounted to persecution because the state had not taken any steps to eliminate the practice. Further, if the religion is intertwined with the state, there may be an overlap between religious and political persecution and the state may play an active role in the persecution. Circumstances where such an overlap is likely include where the state is a theocracy, where the state supports or favours a particular religion or where religious groups undertake political activities. For example, in Refugee Appeal No 71427/99, Haines QC held that the claimant, a woman who objected to the inequality of women’s rights in marriage, divorce and custody in Iran, had a well-founded fear by reason of her religion because Iran was an Islamic theocratic state where religion was used to justify the treatment of women in its laws.
The Refugee Convention definition of a refugee makes no formal distinction between male and female refugees. However, the criteria for determining refugee status have been interpreted in a way that has marginalised the experiences of women. This is because women often suffer persecution in the private sphere by non-state actors or ‘mere discrimination’ as opposed to persecution. The decision in Khawar helps to challenge gendered public/private distinctions in international refugee law by recognising that a state has a public duty to protect its citizens from private harms. However, the facts and decision in Khawar evidence some limitations that leave room for further developments in refugee case law. Many issues remain unresolved. What forms of gender-specific harms may amount to serious harm or persecution? This question is particularly important when dealing with discriminatory laws against women, such as limitations on their rights to work and be educated. If states do have a duty to protect their citizens from harm by non-state actors, what is the measure of this duty? What happens where a state has tried to protect a woman from domestic violence but has ultimately been unsuccessful? Should women be able to claim refugee status for gender-related persecution for reasons of political opinion, religion or their membership of a particular social group? I have argued that women often suffer persecution because they are women and that private harms, such as domestic violence, are not necessarily committed for personal motivations. Thus the decision in Khawar represents an important step towards interpreting the refugee definition in a more gender-sensitive way, but there is still more work to be done.
[∗] BA/LLB (Hons), Australian National University; LLM candidate, New York University. I wish to thank Professor Hilary Charlesworth and two anonymous referees for their suggestions on earlier drafts of this article.
 The Convention Relating to the Status of Refugees done at Geneva on 28 July 1951, as amended by the Protocol relating to the Status of Refugees done at New York on 31 January 1967, art 1A(2).
 H Crawley, Refugees and Gender: Law and Process (2001) 21; T Spijkerboer, Gender and Refugee Status (2000) 65.
 Eg, Minister for Immigration and Multicultural Affairs v Haji Ibrahim  HCA 55; (2000) 204 CLR 1; Chen Shi Hai v Minister for Immigration and Multicultural Affairs  HCA 19; (2000) 201 CLR 293; Applicant A v Minister for Immigration and Ethnic Affairs (1997) 190 CLR 225; Chan v Minister for Immigration and Ethnic Affairs (1989) 169 CLR 379.
 Eg, D Indra, ‘Gender: A Key Dimension of the Refugee Experience’ (1987) 6 Refuge 3; N Kelly, ‘Gender-Related Persecution: Assessing the Asylum Claims of Women’ (1993) 26 Cornell International Law Journal 625; E Love, ‘Equality in Political Asylum Law: For a Legislative Recognition of Gender-Based Persecution’ (1994) 17 Harvard Women’s Law Journal 133; A Macklin, ‘Refugee Women and the Imperative of Categories’ (1995) 17 Human Rights Quarterly 213; D Anker, L Gilbert and N Kelly, ‘Women Whose Governments are Unable or Unwilling to Provide Reasonable Protection from Domestic Violence May Qualify as Refugees Under United States Asylum Law’ (1997) 11 Georgetown Immigration Law Journal 709; Crawley, above n 2.
 (2002) 187 ALR 574.
 H Charlesworth and C Chinkin, The Boundaries of International Law: A Feminist Analysis (2000) 56.
 In this context, ‘sex’ refers to biological differences between men and women, while ‘gender’ refers to social distinctions based on biological sex, including constructions about the identity, status and roles of women and men: see Crawley, above n 2, 6-7.
 K O’Donovan, Sexual Divisions in Law (1985) 2-3.
 M Thornton (ed), Public and Private: Feminist Legal Debates (1995) 11; K O’Donovan, ibid, 12.
 O’Donovan, ibid, 3-4.
 C Hanisch, ‘The Personal is Political’ in J Agel (ed), The Radical Therapist (1971); Thornton, above n 9, 9 and 16; O’Donovan, ibid, 18.
 H Charlesworth, C Chinkin and S Wright, ‘Feminist Approaches to International Law’ (1991) 85 American Journal of International Law 613, 625-34.
 Crawley, above n 2, 129.
 H Charlesworth, ‘Worlds Apart: Public/Private Distinctions in International Law’ in Thornton (ed), above n 9, 243, 244-45.
 Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment (10 December 1984), 1465 UNTS 85, art 1(1).
 Charlesworth, Chinkin and Wright, above n 12, 628-9; Charlesworth and Chinkin, above n 6, 234-35.
 Charlesworth and Chinkin, above n 6, 206-08, 233-37.
 C MacKinnon, ‘Crimes of War, Crimes of Peace’ (1993) 37 UCLA Women’s Law Journal 59, 69-70.
 Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW) (18 December 1979), 1249 UNTS 13, art 2; Universal Declaration of Human Rights, GA Res 217A (1948), art 2; International Covenant on Civil and Political Rights (16 December 1966), 999 UNTS 171, arts 2 and 3; International Covenant on Economic, Social and Cultural Rights (16 December 1966), 993 UNTS 3, arts 2 and 3.
 Eg, CEDAW has much weaker implementation standards than other international human rights instruments, such as the International Covenant on Civil and Political Rights and the Convention on the Elimination of All Forms of Racial Discrimination (21 December 1965), 660 UNTS 195. In stark contrast to the few substantive reservations made to the Convention on the Elimination of All Forms of Racial Discrimination, many states made reservations or declarations of understanding when becoming parties to CEDAW: see Charlesworth, Chinkin and Wright, above n 12, 632-4.
 H Steiner and P Alston, International Human Rights in Context (2000) 158-59.
 C Bunch, ‘Women’s Rights as Human Rights: Towards a Re-vision of Human Rights’ (1990) 12 Human Rights Quarterly 486; Immigration and Naturalisation Service, Considerations for Asylum Officers Adjudicating Asylum Claims from Women (1995) 2; D Anker, Law of Asylum in the United States (3rd ed, 1999) 253; Anker, Gilbert and Kelly, above n 4, 711.
 CEDAW, General Recommendation No 12 of CEDAW Committee, 8th Session 1989, UN Doc HRI \GEN\1\Rev.1, 78 (1994).
 CEDAW, General Recommendation No 19 of CEDAW Committee, 11th Session 1992, UN Doc HRI \GEN\1\Rev.1, 84 (1994).
 UN Doc A/48/629; (1994) 33 ILM 1050.
 For an analysis of the similarities, differences and crossover between human rights and refugee law, particularly with respect to gender, see D Anker, ‘Refugee Law, Gender, and the Human Rights Paradigm’ (2002) 15 Harvard Human Rights Journal 133.
 Convention Relating to the Status of Refugees (28 July 1951), 189 UNTS 150 art 33(1).
 The full text of the definition in art 1A(2) of the Refugee Convention provides that a refugee is a person who ‘owing to a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion, is outside the country of his nationality and is unable or, owing to such fear, is unwilling to avail himself of the protection of that country; or who, not having a nationality and being outside the country of his former habitual residence, is unable, or owing to such fear, is unwilling to return to it’.
 For an alternative critique, see J Greatbatch, ‘The Gender Difference: Feminist Critiques of Refugee Discourse’ (1989) 1 International Journal of Refugee Law 518, 520 (arguing that public/private distinctions have been overemphasised in refugee law) and K Engle ‘After the Collapse of the Public/Private Distinctions: Strategizing Women’s Rights’ in D Dallmeyer (ed), Reconceiving Reality: Women and International Law (1993) 143 (arguing that keeping the private sphere unregulated is not necessarily bad for women because it can be used to protect women’s rights such as the right to abortion).
 Macklin, above n 4; Anker, Gilbert and Kelly, above n 4; Anker, above n 22; Crawley, above n 2; R Bacon and K Booth, ‘The Intersection of Refugee Law and Gender: Private Harm & Public Responsibility – Islam; Ex parte Shah Examined’  UNSWLawJl 57; (2000) 23 University of New South Wales Law Journal 135, 137.
 Some commentators distinguish between women being persecuted as women and women being persecuted because they are women. The problem with this distinction is that it appears to limit persecution of women as women to persecution that uses gender-specific means, rather than also including persecution that has a gender-specific outcome. See Crawley, above n 2, 6-9; Macklin, above n 4, 258-9; Kelly, above n 4, 625.
 For other authors who make the same distinction, see United Nations Human Rights Commission, Report of the Special Rapporteur on Violence Against Women, its Causes and Consequences (1995) 7, 14; Immigration and Nationality Service, above n 22, 4; Immigration Appellate Authority, Asylum Gender Guidelines (2000) [1.13]-[1.15] and [2A.17]; Anker, above n 22, 252-253; and Anker, Gilbert and Kelly, above n 4, 713.
 In some cases, gender-specific harm is unique to one gender because it is perpetrated by gender-specific means, eg, vaginal rape or female genital mutilation. In other cases, the form of harm is not gender-specific but its outcome is. For example, domestic violence in the form of slapping, beating and kicking could be perpetrated against men and women alike. However, domestic violence is still a gender-specific harm because it is far more likely to involve a husband physically abusing his wife than a wife physically abusing her husband.
 Refugee Women’s Legal Group, Gender Guidelines for the Determination of Asylum Claims in the UK (1998) [1.11]-[1.15]; Immigration Appellate Authority, above n 32; Crawley, above n 2, 8; Macklin, above n 4, 259.
 UNHCR, Handbook on Procedures and Criteria for Determining Refugee Status (Redited 1992) . See also s 91R(2) of the Migration Act 1958 (Cth), which provides that ‘serious harm’ includes threats to life or liberty, significant physical harassment or ill-treatment, or significant economic hardship, denial of basic services or denial of capacity to earn a living of any kind where that would threaten a person’s capacity to subsist.
 Canada (Attorney-General) v Ward  2 SCR 689; see also J Hathaway, The Law of Refugee Status (1991) 112.
 Migration Act 1958 (Cth), s 91R(1)(c).
 Such as serious restrictions on the right to earn a livelihood, to practise religion or to access normally available educational facilities: see UNHCR, above n 35, -.
 UNHCR, above n 35, .
 M Crock, Immigration and Refugee Law in Australia (1998) 148.
 UNHCR, above n 35, 16; UNHCR EXCOM Conclusion No 73 (XLIV) ‘Refugee Protection and Sexual Violence’ (1993); Australia, Department of Immigration and Multicultural Affairs, Refugee and Humanitarian Visa Applications: Guidelines on Gender Issues for Decision-makers (1996) [4.6].
 Eg, Minister for Immigration and Multicultural Affairs v Ndege  FCA 783; (1999) 59 ALD 758.
 Eg, in Kenya, married women are legally required to obtain the consent of their husbands before obtaining a national identity card or passport. Similarly, unmarried women or women under the age of 21 must obtain permission from their fathers to obtain a passport or, if they are married, they must obtain permission from their husbands: see United States Department of State, 1999 Country Reports on Human Rights Practices (2000).
 Eg, Hathaway distinguishes four levels of rights. The first includes core civil and political rights, such as the right to freedom from arbitrary deprivation of life, which are non-derogable in any circumstances so all violations amount to persecution. The second includes other civil and political rights, such as the right to a fair trial, which a state may derogate from during an officially proclaimed public emergency but otherwise violations amount to persecution. The third includes positive duties on the state in the form of economic and social rights, such as the right to work and to receive basic education, which are unenforceable but a state should seek to realise within the limits of available resources and in a non-discriminatory way. There may be persecution if these rights are implemented in either a discriminatory way or the failure to implement these rights is not grounded in the absolute lack of resources. The fourth level is hortatory rights, such as the right to own private property and to be protected against unemployment, which are simply goals rather than enforceable obligations. See Hathaway, above n 36, 108-11.
 Charlesworth and Chinkin, above n 6, 206.
 Eg, Fatin v Immigration and Naturalisation Service 12 F.3d, 1241 (3rd Cir, 1993), where the Court found that gender-specific laws, such as women having to wear a veil, were not sufficient to constitute persecution, even where the female asylum seeker objected to the laws on feminist grounds. See also Fisher v Immigration and Naturalisation Service  USCA9 964; 79 F.3d, 955 (9th Cir, 1996), where the Court found that dress restrictions on Iranian women constituted sex discrimination but did not amount to persecution.
 S Wright, ‘Economic Rights and Social Justice: A Feminist Analysis of Some International Human Rights Conventions  AUYrBkIntLaw 11; (1992) 12 Aust YBIL 241, 249.
 UNHCR, above n 35, . See also Horvath v Secretary of State for the Home Department  3 WLR 379 (HL).
 M Randall, ‘Refugee Law and State Accountability for Violence Against Women: A Comparative Analysis of Legal Approaches to Recognizing Asylum Claims Based on Gender Persecution’ (2002) 25 Harvard Women’s Law Journal 281, 301-5; M Stevens, ‘Recognizing Gender-Specific Persecution: A Proposal to Add Gender as a Sixth Refugee Category’ (1993) 3 Cornell Journal of Law & Public Policy 179.
 Macklin, above n 4, 257. Harvey argues that the recent attempts to make the Refugee Convention gender-sensitive have led to interpretations that should be recognised as rational reconstructions of the law: C Harvey, ‘Gender, Refugee Law and the Politics of Interpretation’ (2001) 12 International Journal of Refugee Law 680, 683.
 Anker, above n 26, 139.
 Refugee Women’s Legal Group, above n 31, [1.3].
 UNHCR, Guidelines on the Protection of Refugee Women (1991) 36; J Castel, ‘Rape: Sexual Assault and the Meaning of Persecution’ (1992) 4 International Journal of Refugee Law 39; Crawley, above n 2, 50; Kelly, above n 4, 646.
 Eg, above n 42.
 See generally, Crawley, above n 2, 21-6, 79-90.
 Eg, UNHCR, above n 53; UNHCR, Sexual Violence Against Refugees; Guidelines on Prevention and Response (1995); Canadian Immigration and Refugee Board, Guidelines on Women Refugee Claimants Fearing Gender-Related Persecution (1996); Canadian Immigration and Refugee Board, Guidelines on Women Refugee Claimants Fearing Gender-Related Persecution: UPDATE (1996); Immigration and Naturalisation Service, above n 22; Immigration Appellate Authority, Asylum Gender Guidelines (2000).
 Australian Department of Immigration and Multicultural Affairs, above n 41.
  2 SCR 689.
 Ibid 734.
 Ibid 739.
 Chan v Canada (Minister for Employment and Immigration)  3 SCR 593, 642.
 (1992) 97 DLR (4th) 729 (FCAD).
 Ibid 739.
 Ibid 739. This decision was criticised by McHugh J on the basis that, even following a very liberal interpretation of the refugee definition, it is difficult to see how this group was a particular social group and even if it was, it does not follow that the applicant was abused because of her membership of that group: Applicant A v Minister for Immigration and Ethnic Affairs, above n 3, 262-63, fn 148.
 (1985) 19 I & N 211.
  USCA9 2002; 801 F.2d 1571 (9th Cir, 1986).
 The Court did not refer to In re Acosta, which was handed down shortly before its decision in Sanchez-Trujillo.
 Applicant A v Minister for Immigration and Ethnic Affairs above n 3, 261, fn 127 per McHugh J; Mayers and Marcel v Minister of Employment and Immigration (Canada) (1992) 97 DLR (4th) 729, 735 per Mahoney JA; Hathaway, above n 36, 161, fn 182; Anker, above n 22, 382. The decision also seems contradictory because it gives the family as a prototypical example of a particular social group despite emphasising voluntary associational relationships.
  USCA9 421; 225 F.3d 1084 (9th Cir, 2000).
 Ibid 1093. The Court thus found that gay men with female sexual identities in Mexico could constitute a particular social group because sexual identity is an immutable, fundamental characteristic that is inherent to a person’s identity. See A Bowles, ‘Current Event: Hernandez-Montiel v INS’ (2001) 9 American University Journal of Gender, Social Policy & the Law 717; ‘Recent Case: Immigration Law – Asylum – Ninth Circuit Holds That Persecuted Homosexual Mexican Man with a Female Sexual Identity Qualifies for Asylum Under Particular Social Group Standard. – Hernandez-Montiel v. INS,  USCA9 421; 225 F.3d 1084 (9th Cir, 2000)’ (2001) 114 Harvard Law Review 2569; J Cox, ‘Redefining Gender: Hernandez-Montiel v. INS’ (2001) 24 Houston Journal of International Law 187.
 (1996) 21 I & N 357.
  USCA3 1406; 12 F.3d 1233 (3rd Cir, 1993).
 (1999) Int Dec 3404.
 (2000) Int Dec 3433.
  UKHL 20;  2 AC 629.
 Ibid 645.
 Ibid 659.
 Ibid 635.
 Ibid 653-54.
 Ibid 663.
 Refugee Appeal No 71427/99 (unreported, New Zealand Refugee Status Appeals Authority, 16 August 2000, decision of R P G Haines QC and L Tremewan).
 Khawar v Minister for Immigration and Multicultural Affairs  FCA 1529; (1999) 168 ALR 190, 198.
 Minister for Immigration and Multicultural Affairs v Khawar  FCA 1130; (2000) 101 FCR 501.
 Based on Chen Shi Hai v Minister for Immigration and Multicultural Affairs, above n 3.
 Based on Islam v Secretary of State for the Home Department; R v Immigration Appeal Tribunal; Ex parte Shah, above n 75.
 This is evident from examining the history of the Refugee Convention: see W Kälin, ‘Non-State Agents of Persecution and the Inability of the State to Protect’ (2001) 15 Georgetown Immigration Law Journal 415, 426; A Fortin, ‘The Meaning of “Protection” in the Refugee Definition’ (2001) 12 International Journal of Refugee Law 548; A Grahl-Madsen, The Status of Refugees in International Law (1966) vol 1, 255 and 261. See also the immediate origins of the provision in the 1946 Constitution of the International Refugee Organisation, which viewed a refugee as a person having no consular or diplomatic mission to turn to: ATS 1948, No 16, signed by Australia on 13 May 1947 with effect from 20 August 1948.
 The Macquarie Dictionary, 3rd ed (1997) 1601; see also The Macquarie Dictionary Federation Edition (2001) vol 2, 1423.
 Above n 3. Chen was a child born in contravention of China’s one-child policy. These children, known as ‘black children’, are subjected to discriminatory denials of education, food and healthcare.
 Immigration and Multicultural Affairs v Khawar, above n 5, 591-92 [74-5] per McHugh and Gummow JJ stating that the issue was also left open in Minister for Immigration and Multicultural Affairs v Haji Ibrahim above n 3, 53-5 per Gummow J; 80-1 per Callinan J.
 Chan v Minister for Immigration and Ethnic Affairs, above n 3, 388.
 Immigration and Multicultural Affairs v Khawar, above n 5, 594  per McHugh and Gummow JJ.
 Eg, the High Court has rejected the use of dictionaries to import notions of enmity and malign intention on the part of the state into the concept of persecution: Chen Shi Hai v Minister for Immigration and Multicultural Affairs, 304 .
 Islam v Secretary of State for the Home Department; R v Immigration Appeal Tribunal; Ex parte Shah, above n 75, 653 per Hoffman LJ; see also Refugee Women’s Legal Group, above n 34, [1.17].
 Minister for Immigration and Multicultural Affairs v Haji Ibrahim, above n 3, 50  per Gummow J.
 While Gleeson CJ did not give an example of this point, a relevant example might be the persecution of black South Africans under apartheid.
 Above n 3.
 Eg left-handed people are not a particular social group, but if they were persecuted because they were left-handed, they would no doubt quickly become recognisable in their society as a particular social group. Nonetheless, it would be the attribute of being left-handed that would create the public perception that they were a particular social group, not the persecutory acts: see Applicant A above n 3, 264 per McHugh J.
 Anker, above n 22, 253; Castel, above n 53, 43.
 Above n 42.
  USCA5 306; 809 F.2d 285 (5th Cir, 1987).
 R Bacon and K Booth, above n 30, 137.
 See P Goldberg, ‘Anyplace but Home: Asylum in the United States for Women Fleeing Intimate Violence’ (1993) 26 Cornell International Law Journal 565, 584-85; Anker, above n 22, 193-5; Anker, Gilbert and Kelly, above n 4, 773-2.
 4 Inter. Am. Ct. HR. Ser c. No 4 (1988). This case concerned the disappearance and torture of a Honduran student, where the government was accused of tolerating a practice of kidnapping and disappearances.
 GA Res 48/104, 48 UN GAOR Supp (No 49) 217, UN Doc A/48/49 (1993); (1994) 33 ILM 1050.
 P Mathew, ‘Conformity or Persecution: China’s One Child Policy and Refugee Status’  UNSWLawJl 56; (2000) 23 University of New South Wales Law Journal 103, 120.
 Charlesworth and Chinkin, above n 6, 59; see also Engle, above n 29 (arguing that the unregulated private sphere is not necessarily bad for women because it can be used to protect women’s rights such as the right to abortion).
 Australia, Department of Immigration and Multicultural Affairs, above n 41 [4.6].
 Minister for Immigration and Multicultural Affairs v Haji Ibrahim above n 3, 18-19  per McHugh J; cited with approval by McHugh and Gummow JJ in Minister for Immigration and Multicultural Affairs v Khawar, above n 5, 592 .
 Chan v Minister for Immigration and Ethnic Affairs, above n 3, 388; cited with approval by McHugh and Gummow JJ in Minister for Immigration and Multicultural Affairs v Khawar, above n 5, 592 .
 Chen Shi Hai v Minister for Immigration and Multicultural Affairs, above n 3, 303.
 Macklin, above n 4, 244.
 Hathaway, above n 36, 112.
 United Nations High Commission, above n 41, 10; Goldberg, above n 102, 570, 584-5; Anker, above n 22, 196-7; Anker, Gilbert and Kelly, above n 4, 728, 734.
 Islam v Secretary of State for the Home Department; R v Immigration Appeal Tribunal; Ex parte Shah, above n 75, 635.
 Randall, above n 49, 317.
 A Sinha, ‘Domestic Violence and US Asylum Law: Eliminating the “Cultural Hook” for Claims Involving Gender-Related Persecution’ (2001) 76 New York University Law Review 1562; S Razack, ‘Domestic Violence as Gender Persecution: Policing the Borders of Nation, Race, and Gender’ (1995) 8 Canadian Journal of Women & Law 84.
 Macklin, above n 4, 264; Anker, above n 26, 152-3.
 Canada (Attorney-General) v Ward, above n 36.
 Randall, above n 49, 317.
 G Goodwin-Gill, The Refugee in International Law (2nd ed, 1996) 73; Anker, above n 26, 135.
 See Zalzali v Canada (Minister of Employment and Immigration)  3 FC 605 (FCA).
 N Kelly, ‘The Convention Refugee Definition and Gender-Based Persecution: A Decade’s Progress’ (2002) 13 International Journal of Refugee Law 559, 565.
 Crawley, above n 2, 48.
 Immigration Appellate Authority, above n 32, [2B.3].
 See Crawley, above n 2, 51; see also Immigration Appellate Authority, above n 32, [2B.7].
 Above n 71.
 Interim Dec. 3346 at 5-6 (BIA 1998).
 Immigration Appellate Authority, above n 32 [2B.8].
 UNHCR, above n 35, .
 Mathew, above n 105 .
 Canadian Immigration and Refugee Board, above n 56, pt B; see also Macklin, above n 4, 229-232; Crawley, above n 2, 52.
 Above n 3, 258.
 Chen Shi Hai v Minister for Immigration and Multicultural Affairs, above n 3, 301 .
 Islam v Secretary of State for the Home Department; R v Immigration Appeal Tribunal, Ex parte Shah, above n 75, 663.
 P Mathew, ‘Islam v Secretary of State for the Home Department, and R v Immigration Appeal Tribunal, Ex parte Shah  UKHL 20;  2 AC 629’ (2001) 95 American Journal of International Law 671, 676.
 Crawley, above n 2, 58-9.
 Above n 36, 724 cited with approval by Court of Appeal in Horvath v Secretary of State for the Home Department  EWCA Civ 3026;  INLR 15 (CA).
 Minister for Immigration and Multicultural Affairs v Khawar, above n 5, 580-1  per Gleeson CJ, 594  per McHugh and Gummow JJ.
 A Macklin, ‘Cross-Border Shopping for Ideas: A Critical Review of United States, Canadian, and Australian Approaches to Gender-Related Asylum Claims’ (1998) 13 Georgetown Immigration Law Journal 25, 58.
 Eg, Applicant A v Minister for Immigration and Ethnic Affairs, above n 3, 262-63, fn 148 per McHugh J.
 Mathew, above n 135, 675.
 Transcript, lines 365-370.
 Transcript, lines 360-365.
 See generally R Dixon, ‘Rape as a Crime in International Humanitarian Law: Where to from Here?’ (2002) 13 European Journal of International Law 697.
 The Prosecutor v Kunarac, Kovac and Vukovic, (Case No IT-96-23T) Judgment, 22 February 2001.
 Ibid .
 Ibid .
 Prosecutor v Akayesu, (Case No ICTR-96-4-T) Judgment, 2 September 1998.
 C MacKinnon, ‘Rape, Genocide and Women’s Human Rights’ (1994) 17 Harvard Women’s Law Journal 5, 9-10.
 K Crenshaw, ‘Mapping the Margins: Intersectionality, Identity Politics, and Violence Against Women of Color’ (1991) 43 Stanford Law Review 1241; see also Macklin, above n 4, 263.
 L Green, ‘Gender Hate Propaganda and Sexual Violence in the Rwandan Genocide: An Argument for Intersectionality in International Law’ (2002) 33 Columbia Human Rights Law Review 733.
 Prosecutor v Delalic, (Case No IT-96-21) 16 November 1998, -.
 Statute of the International Criminal Court, Rome, 1998, art 7(h); (1998) 37 ILM 999.
 K Patel, ‘Recognizing the Rape of Bosnian Women as Gender-based Persecution’ (1994) 60 Brooklyn Law Review 929, 931-32, 949.
 T Spijkerboer, Women and Refugee Status: Beyond the Public/Private Distinctions (1994) 26-7; see also G Goodwin-Gill, The Refugee in International Law (2nd ed, 1996) 364.
 Sabalic, who reported that some women had their breasts cut off and stomachs slit open after being raped in Bosnia, concluded that the sexual nature of the violence perpetrated against women was an expression of hatred towards women: see R Seifert, ‘War and Rape: A Preliminary Analysis’ in A Stiglmayer (ed) Mass Rape: The War against Women in Bosnia-Herzegovina, (1994) 54, 65, fn 41. Likewise, Brownmiller argues that women in war are not raped because they ‘belong to the enemy camp, but because they are women and therefore are enemies’: see S Brownmiller, Against Our Will: Men, Women and Rape (1975) 69.
 C MacKinnon, Feminism Unmodified: Discourses on Life and Law (1987) 85-6.
 Randall, above n 49; Stevens, above n 49.
 Macklin, above n 4, 256-73.
 D Nebesar, ‘Gender-Based Violence as a Weapon of War’ (1998) 4 University of California International Law & Policy 147, 149-150; J Campanaro, ‘Women, War, and International Law: The Historical Treatment of Gender-Based War Crimes’ (2001) 89 Georgetown Law Journal 2557; R Seifert, ‘War and Rape: A Preliminary Analysis’ in A Stiglmayer (ed), Mass Rape: The War against Women in Bosnia-Herzegovina (1994) 54.
 A Cahill, Rethinking Rape (2001) 3.
 Bunch, above n 22, 490-91.
 Above n 81.
 Ibid .
  USCA9 557; 813 F.2d 1432 (9th Cir, 1987).
 Mathew, above n 135, 674.
 Macklin, above n 4, 246.
 Randall, above n 49, 292-93; Macklin, , above n 4, 245.
 Macklin gives an extreme Canadian example of this point where a social group was found of ‘new citizens of Israel who are women recently arrived from elements of the former Soviet Union and who are not yet well integrated into Israeli society … who are lured into prostitution and threatened and exploited by individuals not connected to government, and who can demonstrate indifference in their plight by front-line authorities to whom they would normally be expected to turn for protection’: see Macklin, above n 139, 60-1 quoting from Litvinov v Canada  FTR 60.
 UNHCR EXCOM Conclusion No 39 (XXXVI) ‘Refugee Women and International Protection’ (1985), (k).
 N Kelly, above n 4, 657-58 (distinguishing between the wider social group and the narrower refugee group within that social group); Castel, above n 53, 53; Macklin, above n 4, 247-48.
 Anker, above n 22, 393.
 Randall, above n 49, 286-87.
 Australian Department of Immigration and Multicultural Affairs, above n 41, [2.10].
 M Crock, Immigration and Refugee Law in Australia (1998) 149.
 See Immigration and Refugee Board, Gender-Related Claims, 1993/4-1997/8, see Macklin, above n 139, 34.
 Spijkerboer, above n 155, 90.
 Macklin, above n 4, 241; Crawley, above n 2, 21.
 Crawley, above n 2, 82-4; Macklin, above n 4, 240-1.
 Spijkerboer, above n 155, 57.
 Greatbatch, above n 29.
 Above n 72, 1241.
 See also Immigration Appellate Authority, above n 32, [3.25]; N Kelly, above n 4, 637-38; Anker, above n 22, 370-72; Anker, Gilbert and Kelly, above n 4, 741-42.
 Spijkerboer, above n 2, 46; Canadian Immigration and Refugee Board, above n 56, [4.16]; Sinha, above n 116, 1593-94.
 Refugee Review Tribunal, N97/05699 (21 July 1997) (Hudson).
 Macklin, above n 139, 45.
 UNHCR, above n 35, -.
 Crawley, above n 2, 84-5.
  USCA9 964; 79 F.3d 955 (9th Cir, 1996).
 Mathew, above n 105, 123.
 Above n 165.
 Macklin, above n 139, 58-9; Anker, above n 22, 372.
  USCA5 306; 809 F.2d 285 (5th Cir, 1987)
 Immigration and Naturalisation Service, above n 22, pt IIIA.
 Greatbatch, above n 29, 521; Anker, Gilbert and Kelly, above n 4, 743.
 Canadian Immigration and Refugee Board, above n 56, pt A (II); Immigration Appellate Authority, above n 32, [3.12].
 S Akram, ‘Orientalism Revisited in Asylum and Refugee Claims’ (2000) 12 International Journal of Refugee Law 7.
 Australian Department of Immigration and Multicultural Affairs, above n 53, [4.30]; see also Immigration Appellate Authority, above n 32, [3.13].
 See generally T Higgins, ‘Anti-essentialism, Relativism, and Human Rights’ (1996) 19 Harvard Women’s Law Journal 89.
 Anker, above n 26, 71.
 Above n 5, 581  per Gleeson CJ.
 Eg, arts 2(f) and 5(a) of CEDAW require states parties to take all appropriate measures, including legislation, to modify or abolish laws, customs and practices that constitute discrimination against women, and to modify social and cultural patterns of conduct of men and women with a view to eliminating prejudices and customary practices based on the idea of the inferiority of women. Similarly, the Declaration on the Elimination of Violence against Women GA Res 48/104, 48 UN GAOR Supp (No 49) 217, UN Doc A/48/49 (1993) art 4 provides that ‘States should condemn violence against women and should not invoke any custom, tradition or religious consideration to avoid their obligations’.
 Anker, above n 26, 71.
 Above n 71.
 Immigration Appellate Authority, above n 32, [3.15].
 Above n 81.