Inaugural Meeting IARLJ Australia/New Zealand Chapter - 10 March 2000
The Hon Justice Deidre O'Connor, Federal Court of Australia 




1.    Private vs Public Harm
2.    "Women" as a particular social group
3.    Defining the group by the persecution feared






[1] Thank you for the opportunity to address the Association's Chapter Conference. I am mindful of the significant role of organisations such as this one in contributing to the development of best practice in the domestic implementation of our countries' international obligations. I hope that we will take with us from the conference a collection of ideas and new perspectives and a renewed commitment with which to implement them.

[2] In October 1998 the Association's Working Party on Particular Social Group submitted a discussion paper on what is a thorny and important question - the ambit of the fifth enumerated ground of persecution in the Refugees Convention. What I propose to do this afternoon is to pick up on several of the issues raised by that paper and to demonstrate some ways in which the consideration of gender creates further problems in what is already a difficult area of refugee law. Initially, I will give a brief overview of various approaches taken to the "particular social group" ground across different jurisdictions, with particular reference to gender-based persecution. I will then focus on the recent House of Lords decision, Islam; Ex parte Shah which deals with some interesting questions about the interface between gender and refugee law in a contemporary setting, perhaps one might say for the new millennium.


[3] Is the definition of "refugee" contained in Article 1A(2) really as inclusive as it first appears? In particular, what is the significance of the omission of "gender" as a ground of persecution?
[4] As we know, the Refugees Convention was drafted in 1950, at a time when problems occurring in the aftermath of World War II were uppermost in the minds of the drafters. With the exception of the "particular social group" ground, the 1951 Convention lists the most apparent forms of discrimination then known; that is, large groups covered by race, religion and political opinion.1 Even the ground of "particular social group" was conceived in response to contemporary problems such as the displacement of the land-owning classes in Communist countries.

[5] The changing character of groups of refugees, states which are sources of conflict, and increased understanding and acceptance of women's issues as human rights issues all combine to highlight the importance of taking an evolutionary approach to the interpretation of the Refugees Convention. As Lord Hope pointed out in Islam; Ex parte Shah; an evolutionary approach:

[6] The adoption of an evolutionary approach in different jurisdictions has enabled some decision-makers to take a more 'gender-sensitive' approach to the interpretation and application of refugee law. It has tended regrettably to happen at the expense of consistency.

[7] One of the values of a forum such as this one is the opportunity it gives us to explore how decision makers in other jurisdictions are approaching common problems, while at the same time encouraging a global approach to issues such as women's rights, human rights and refugee status.


[8] The types of difficult issues which arise when gender and refugee law intersect have been described by Justice Moore in Faddoul.3

[9] The issues Justice Moore pinpoints have only become issues since gender-aware perspectives and ideas, such as an understanding of the public/private dichotomy and the way in which it creates "private" harm, have been incorporated into contemporary thinking.

[10] Many claimants from gender-based groups who, if their stories are believed are undoubtedly persecuted, are unsuccessful. Decisions in this area demonstrate that decision makers are wrangling with issues that arise from characterising gender-based claims as "particular social group" claims. The decisions discuss "floodgates" arguments, characterise persecution, define the relevant "social group", and establish a Convention nexus between harm feared and membership of a group. However, a general recognition of the types of issues which arise when gender and refugee law intersect has not resulted in any clear or consistent approach to negotiating these issues when they arise in practice.

[11] Why are gender-based claims problematic? Some answers to this question can be found in the arguments raised by critical refugee law scholars concerning the treatment of women within the refugee context - that women's experiences are not recognised in the international sphere of public actors; that economic and social status relegate women to the private sphere of home life, not political life. As a result, their experiences occur within that sphere. Domestic violence, rape, forced marriage, genital mutilation, these are issues which, for a long time, have either gone unrecognised or been excluded from the range of matters which the international community accepts it can influence. 'Invisibility' is an apt term to describe the historical plight of women's human rights. Given this invisibility, once recognition of women's issues does occur, it is then up to the existing system to respond.

[12] How is it possible that a particular social group might be formulated, given the principle that a particular social group cannot be defined by reference to the feared persecutory conduct? When applied to the types of problems commonly faced by women, for example, women victims of domestic violence, claimants face the difficulty of characterising both the relevant group and the relevant persecution. To define the group as "women victims of domestic violence", for example, defines the group by the persecution feared. On the other hand, to define the relevant group as simply "women", risks the criticism that the "group" is too broad and, in terms of Australian law, is not sufficiently cognisable or set apart from the rest of society.5 In terms of certain North American case law, expressed in Sanchez-Trujillo, such a group is arguably not homogenous or cohesive enough to constitute a particular social group.6 Such a widely defined group raises concerns if one uses a purely policy perspective, in that if claimants are granted refugee status on the basis of membership of a very large group, arguably this will open the 'floodgates' for applications from other members of the group.

[13] Different jurisdictions throughout the world are developing different ways of handling these difficulties, and decision makers are employing a range of legal, conceptual and policy solutions.


[14] As concluded by Carlier in his book Who is a Refugee? A Comparative Case Law Study,7 of the fifteen countries in that study, the main themes in defining particular social group congregate along the lines of common social characteristics,8 common immutable characteristics,9 innate and unchangeable characteristics10 and, more specifically, groups with a homogenous background, behaviour and social status.11 While in Australia and Canada the approaches seek a formula, in Germany decisions are reached on an "intuitive sense of persecution" based on membership of a particular social group which is derived from stable and inalterable characteristics. The Ward decision from Canada enunciated a working role rather than "an unyielding deterministic approach to resolving whether a refugee claimant could be classified within a particular social group."12

[15] The paramount consideration in determining whether a particular social group exists is the "general underlying themes of the defence of human rights and anti-discrimination."13 Despite these matters in common, the application of the principles concerning the "particular social group" ground of the Convention to cases involving gender issues produces inconsistencies that are not necessarily attributable to factual differences between like cases. As a fact-driven area of the law, one might hope to explain a number of seeming inconsistencies by reference to the claimant's circumstances. There are examples of differing applications of the law to the facts of similar cases. These variations occur across jurisdictions and also, perhaps more alarmingly, within domestic jurisdictions. Three examples of this are:

l.    differences in the degree to which traditionally "private" harms are recognised as public and therefore within the ambit of the Convention;

2.    different answers to the question whether the category of "women" is capable of constituting a "particular social group" within the meaning of the Convention;

3.    differing approaches to avoiding the circularity of defining a group by the persecution feared.

1.    Private vs Public Harm

[16] Firstly, there are differences in the degree to which traditionally "private" harms are recognised as public and therefore within the ambit of the Convention. In an Australian context, this is demonstrated by two recent decisions of the Federal Court of Australia, Ndege and Khawar. Both cases involved women victims of domestic violence from societies in which women occupied low social status, domestic violence was rife and there were no support structures in place for women victims of spousal abuse. Indeed, in Khawar it was noted that domestic violence was officially tolerated in that no state assistance was forthcoming despite the victim's repeated requests for help.

[17] In Ndege, the Tribunal found that there was a real chance the claimant would suffer serious harm at the hands of her husband if she returned and that such harm would constitute persecution for a Convention reason, namely, for reasons of her membership of the particular social group "married women in Tanzania". The Tribunal found there was a direct nexus between her membership of the group and the motivation of the state to persecute by denying the claimant access to protection. On appeal to the Federal Court for judicial review, however, Justice Weinberg overturned the TribunaI's decision and held that the Tribunal's conclusion that the state was the source of the Convention-related persecution was not properly open to it. The Court conceptualised the harm as private harm, requiring the perpetrator of the violence to be motivated by a Convention reason in order to establish a nexus with the Convention.

[18] This can be contrasted with the approach taken by Justice Branson in Khawar and with at least two members of the House of Lords in Islam; Ex parte Shah.
[19] In Khawar, Justice Branson suggests that the refusal or failure of State law enforcement officers to take steps to protect members of a particular social group from violence is itself capable of amounting to persecution under the Convention. She found that the motivation of the claimant's husband to harm her would have no relevance in such a case.

[20] In Islam; Ex parte Shah, it was said that:

[21] It should be noted that the cases are unclear as to identification of the relevant persecution, for example, whether the persecutor is the violent husband, or the state; or whether the battering or the refusal of state protection constitutes the relevant persecution. This complication is demonstrated in Islam; Ex parte Shah, which I will discuss in more detail in a moment.

[22] The approaches taken in Khawar and Islam; Ex parte Shah demonstrate a willingness to look beyond the acts of the perpetrator and take a broader, more contextualised approach to assessing an applicant's claims. The decision makers in these cases looked at the claimant's individual stories, but did so within the social and power context in which these women lived. It is arguable that the cases conceptualise the persecution as the withholding of state protection, thereby positing the harm feared by women squarely in the public realm. As analysed by Lord Hoffman in Islam; Ex parte Shah:

[23] This approach stands in stark contrast to one which goes no further than conceptualising domestic violence as private harm, arising out of a personal relationship.

2.    "Women" as a particular social group

[24] Secondly, there are differing views over whether the category of "women" is capable of constituting a "particular social group" within the meaning of the Convention. While the courts in Australia have not definitively decided this issue, differently constituted Tribunals have come to opposite conclusions in cases involving the same countries and similar fact scenarios. A number of Refugee Review Tribunal decisions have rejected the idea that "women" in a particular society may constitute a particular social group,16 and many more have sought to decide cases on other grounds.17 On the other hand, in cases such as RRT Reference N93/000656 (Aug 3, 1994) the Tribunal has found that "women" can constitute a particular social group. In that case, the decision maker focused on elements such as women's social role as mothers and home makers and their biological characteristics. It concluded that women were set apart and cognisable as a group within the society in question.

[25] Various reasons are given for accepting or rejecting the suggestion that "women" can constitute a particular social group. The argument may be made that "women" is too broad a category to be considered a particular social group or to be cognisable in a society, and that if one woman is accepted on this basis, a flood of applications will be made.18 On the other hand, them are indications in Australian case law that the size of the group is not determinative. Similarly, in Canada, the Canadian Guidelines for decision makers concerning matters of gender state that the size of the particular social group is irrelevant. By way of comparison, race, religion, nationality and political opinion are also characteristics which are shared by large numbers of people.19

[26] In Chan v Canada the dissentient rejected the floodgates argument as a relevant factor, stating that "I am mindful that the possibility of a flood of refugees might be a legitimate political concern, but it is not an appropriate legal consideration".20 This view is consistent with arguments that refugee status is an individual remedy; even if a Convention nexus is established, this may not be sufficient to give rise to refugee status. Other elements of the refugee definition, such as the need to prove objective fear of future persecution, the seriousness of such persecution, and the lack of internal flight alternatives, may prevent female claimants from attaining refugee status. Further, lack of resources to flee their country of origin often precludes women from constituting a 'massive influx' of refugee claimants.

3.    Defining the group by the persecution feared

[27] Thirdly, decision-makers in various jurisdictions differ in the extent to which they attempt to avoid the circularity of defining a particular social group by reference to the persecution feared. The clear statement of principle laid down in Applicant A, that the persecution cannot be the sole defining characteristic of the group in question, has been quoted and applied in several overseas judgments, including the recent House of Lords decision in Islam; Ex parte Shah.21 The application of this principle would preclude a decision maker from granting refugee status to victims of domestic violence or other types of gender-based harm on the basis that they were members of a particular social group constituted by, for instance, "women who are victims of domestic violence". This principle has been heeded by decision makers in the Australian jurisdiction; decision makers who have granted refugee status to women on the basis of their membership of a particular social group have generally framed the relevant group in such a way that it was not defined by reference to the feared persecutory conduct.22
[28] In the Canadian jurisdiction, on the other hand, there are numerous cases in which particular social groups have been defined by reference to the harm feared by a claimant.

[29] Finally, all of these approaches may be contrasted with the approach commonly adopted in the US, where gender-based claims are often linked to the ground of political opinion ground in addition to, or instead of, "particular social group".23


[30] The divergent approaches to claims involving women as a particular social group indicate that these matters raise difficult issues. The fact that different outcomes are reached in like matters, for a range of different reasons, also indicates that judges and decision makers make choices about how best to achieve the humanitarian objectives of the Convention. As Lord Steyn, quoting from Justice Sedley, reiterated in his judgment in Islam; Ex parte Shah:

[31] In the course of such a process, the assumptions, values and preconceptions of individual decision makers inevitably come into play. This is particularly so in matters involving gender-based claims and the "particular social group" ground. Choices often need to be made about whether a number of people with related characteristics can sensibly constitute a particular social group, and whether a claimant's story fits within the intended scope of the Convention.
[32] The range of judicial choices open to refugee law decision makers, and the differing approaches taken, are apparent in the recent House of Lords decision of Islam; Ex parte Shah.


[33] This recent House of Lords decision involved the conjoined appeals of two Pakistani women whose claims for asylum in the United Kingdom had been rejected. The central issue to be decided was whether the appellants could claim to be members of a 'particular social group' within the meaning of Article 1A(2) of the Refugees Convention. Both women in these cases were victims of domestic violence in Pakistan at the hands of their husbands, in a legal and social context where the state was unwilling or unable to offer protection. There was a great deal of evidence before the Law Lords that the status of women in Pakistan was particularly low, and that domestic violence and abuse of women is prevalent in that society. The distinctive feature of these cases was that, in Pakistan, women are unprotected by the State, which partly tolerates and partly sanctions discrimination against women. This element of state protection proved crucial in the Lords' determination of causation.

[34] By a majority of 4 to 1 (Lord Millett dissenting) the House of Lords found that women in Pakistan could constitute a particular social group within the meaning of Article 1A(2).

[35] After examining the history and context in which the Convention was drafted, and the principles referred to in the preamble, two out of three Law Lords (Lord Millett dissenting) held that persecution for reasons of gender was intended to be covered by the phrase "particular social group" in Art 1A(2). Lord Hoffman considered that the inclusion of particular social group as a ground recognised that there might be different criteria for discrimination which would be as offensive to principles of human rights as the forms of abuse covered by the other grounds.25 Lord Millett, on the other hand, thought it noteworthy that sex was included as a basis of discrimination in the Universal Declaration of Human Rights, but not among the grounds in the Refugees Convention.26

[36] While all three judges agreed that the nature of each appellant's experience raised issues relating to the "particular social group" ground of the Convention, Lord Millett differed with Lords Steyn, Hoffman and Hope in his approach to these issues. On the question whether the women belonged to a particular social group, the appellants identified two possible groups, namely, 'women in Pakistan' and 'women in Pakistan who are unprotected by the State and who transgress social mores'. Lords Steyn, Hoffman and Hope agreed that, in the circumstances of these cases, both groups could constitute "particular social groups" within the meaning of the Convention. Lord Millett, on the other hand, considered that the second formulation suggested by the appellants was circular, as it sought to define the group by the persecution feared.

[37] Following on from this, the Lords took differing approaches to the issue of Convention nexus. Lord Millett found that, even if the wider group of "Pakistani women" did constitute a particular social group, the appellants did not fear persecution for reasons of membership of such a group. While recognising that women in Pakistan were discriminated against, his Lordship distinguished this from Convention-related persecution. In his view, the appellants feared harm because they had transgressed social norms, not because they were women. His Lordship stated at 1044:

[38] Lord Millett's approach to the question whether the harm feared by these women was Convention-related differed markedly from that of Lords Steyn and Hoffman. While Lord Millett saw the reason for the harm feared as being punishment for transgressing social norms, Lords Steyn and Hoffman took a more holistic view, taking into account the situation of women in Pakistani society and the attitude of the State. Lord Steyn found that, given the central feature of State-tolerated and state-sanctioned gender discrimination in Pakistan, the argument that the appellants feared persecution not because of their membership of a social group but because of the hostility of their husbands was unrealistic.27 Lord Hoffman considered that, while the husbands' violence was a personal affair directed towards the appellants as individuals, the evidence was that the State would not assist them because they were women. It therefore denied them a protection against violence which it would have given to men.28

[39] The majority's view is illustrated by the following example, given in Lord Hoffman's judgment.

[40] Interestingly, Lord Millett accepted that a State may persecute members of a particular social group by openly withdrawing its protection and leaving them to the mercy of criminal elements, and that the victims in such a situation should be regarded as the victims of official persecution by the State. However, he distinguished those cases involving domestic violence as his Lordship considered that the appellants had failed to prove that Pakistan had withdrawn its protection for a Convention reason.30

[41] Perhaps the major difference between the approaches of the majority and minority lies in their assumptions about gender, and the extent to which each was prepared to conceptualise what is traditionally 'private harm' (domestic violence) as serious harm which should be addressed in the public sphere, and for which the State must bear some responsibility. The majority accepted that, in certain circumstances where a State fails to protect its citizens, private, gender-based harm can constitute persecution within the meaning of the Convention. Their reasoning indicates that the 'persecutor' or 'agent of harm' in these cases was the State, and that the persecution was the act of withholding protection for a Convention reason. This is significant in Australian refugee law, as it is necessary to determine the identity of the persecutor in order to determine why the agent was motivated to persecute claimant. This in turn enables the decision maker to determine whether the persecution is "for reasons of" a Convention ground?31

[42] Lord Millett, on the other hand, considered that the harm suffered by the appellants was restricted to the violence inflicted by the husband. He concluded that the harm feared was therefore personal, not Convention-related, and that the State's discriminatory attitude towards women did not amount to persecution in these cases.

[43] Significantly, Lord Millett commented that "it is difficult to imagine a society in which women are actually subjected to serious harm simply because they are women".32 Perhaps his Lordship was unaware of the many forms of gender-based disadvantage now recognised, such as inability to work; inability to own property; inability to vote; inability to obtain divorce; lower pay for equal work; inability to borrow money; the 'glass ceiling'; sexual harassment; and sexual violence - forms of harm which are imposed on women for no reason other than that they are women. Or perhaps he considered that these harms are simply not serious enough to amount to persecution. Another explanation, and one implicit in his reasoning, is that his Lordship did not conceptualise private forms of harm, such as domestic violence, as harm which should be addressed by the international community alongside long-recognised forms of (public) harm, such as putting people in concentration camps because of their political opinion, religion or race.

[44] Lord Millett's conclusion that gender-based harm was not intended to be covered by the Refugees Convention stands in stark contrast to his view that the kinds of discrimination intended to be protected against under the "particular social group" ground included discrimination on the basis of language, social origin, economic class, ethnicity, birth or other status.33 It is at least arguable that discrimination on the basis of gender has qualities in common with these other bases of discrimination - they are immutable, and they cause the members of that group to be treated differently in many aspects of their lives.


[45] Islam; Ex parte Shah is not a 'floodgates' case. The Law Lords are careful to note at several points in the judgment that the case turns upon its facts - that while the claimant were successful as Pakistani women, it is not to say that all Pakistani women will be successful in future. Each case must be examined on its merits; a successful claim is constituted by a convergence of all elements of the definition and therefore differences in the individual factual matrix will produce a different result.

[46] While the full impact of the decisions are yet to be explored in the judicial reasoning of foreign jurisdictions, the case has prompted much timely public discussion in the UK as to its future interpretation and application. For example, there is some writing on the question of whether the reasoning can extend to other sex or gender-based claims such as gay men persecuted for their sexual orientation.

[47] Apart from throwing up questions as to how we conceptualise the constituent parts of the definition in relation to what constitutes persecution and the agents of harm, the value of Islam; Ex parte Shah to us as decision makers is twofold. Firstly, in a general sense, the case is a good example of the importance of a global approach in this difficult field. The Law Lords sought guidance in jurisprudence from several jurisdictions.

[48] Secondly, the case is important in that it discusses state complicity and its relationship to persecution. The majority decision signals a move towards breaking down the conceptual barriers of public versus private harm in the context of the convention. This is achieved by attributing to the state responsibility for 'private' harm.

[49] Shifting the boundaries of public/private has ramifications for other forms of behaviour which are commonly categorised as private, eg, not only domestic violence, but also forced marriage, female genital mutilation, and serious punishment for transgressing social mores. This represents a move to include these things as types of harm falling within the Convention, as opposed to acts such as torture for political opinion, which have always been acknowledged as falling within the Convention's scope.


[50] Other questions we might ask ourselves but not dealt with here include whether it was the framers' intention for the Convention to provide a response in situations of, for example, forced marriage. Is a woman's refusal to submit to the practice of genital mutilation an issue that the Convention should address? Can a woman's claims be described as "feminist" and thereby invoke the political opinion ground?

[51] More importantly, are the efforts to facilitate the claims of women for refugee status merely artificial constructs called into being to meet the exigencies of the individual case, as alleged by Lord Millett in his dissent in Islam; Ex parte Shah? Should we start again, with a new document specifically targeting the needs of women in international law? Perhaps the Convention itself could be reformed, simply by the addition of gender as a ground for persecution. An additional document focusing on women in the refugee context is problematic - to separate out women recognises the uniqueness of their experience just as it isolates that experience, inhibiting the continuing recognition of women's rights as human rights. A potential problem in creating a place for women outside the refugee status model is that alternative status (for example, a visa granted on humanitarian or provisional grounds) may carry with it less rights or benefits than those which accrue to full refugee status, thereby marginalising women. Perhaps we should address the more difficult task of bringing into the mainstream a gender perspective, incorporating it into existing construction of refugee.34

[52] At its essence, we must ask ourselves whether, at international law, we wish to protect victims of serious violence perpetrated by non-state actors. Can it matter to us if that victim is male or female? What is the best mechanism by which such protection is achieved?

[53] The intersection of gender and refugee law forces us to confront shortcomings in the current structures within which the international community administers aid to displaced persons, in order to fulfil that broad humanitarian purpose. As ever, raising the gender question in turn raises many more questions, often fundamental to the way in which we view our world and its social, legal and economic structures. Increased awareness of the issues and an opportunity to address them may foster an environment in which answers to these questions may be found. It would be better if it were sooner rather than later.

1. Cf Lord Steyn in Islam v Secretary for the Home Department; R v Immigration Appeal Tribunal and Another; Ex parte Shah [1999] 2 WLR 1015 at 1025 (hereafter cited as Islam; Ex parte Shah).
2. Ibid., per Lord Hope at 1038.
3. Faddoul v MIMA [1999] FCA 87.
4. Ibid., at paragraph [23].
5. This principle was laid down by the High Court in Applicant A v Minister for Immigration and Ethnic Affairs and Another (1996) 190 CLR 225, and is neatly summarised by the Full Court in MIMA v Zamora (1998) 51 ALD 1, at 6-7.
6. Sanchez-Trujillo v Immigration and Naturalization Service (1986) 801 F.2d 1571. See also, Islam; Ex parte Shah [1998] 1 WLR 74 at 85.
7. Carlier (ed) Who is a Refugee? A Comparative Case Law Study (Kluwer Law International 1997)
8. This approach is common in Switzerland.
9. The House of Lords in Islam; Ex parte Shah referred approvingly to this phrase, acknowledging the US decision in In Re Acosta (1985) 19 I & N 211. See for instance, Islam; Ex parte Shah per Lord Steyn at 1025.
10. These phrases are common in Canada, as well as in the United Kingdom.
11. These features are looked for in Denmark.
12. As described in Chan v Canada (Minister of Employment and Immigration), [1995] 3 S.C.R 593.
13. Chan v Canada (Minister of Employment and Immigration), [1995] 3 S.C.R. 593 at 642.
14. Islam; Ex parte Shah per Lord Millett at 1045. Lord Steyn made similar remarks in obiter.
15. Islam; Ex parte Shah per Lord Hoffman at 1034 - 1035.
16. See, for instance, RRT Reference N97/15435 (28 January 1998). In this matter the RRT considered that Applicant A precluded a finding that, of themselves, women can constitute a particular social group, as their individual circumstances differ so much in relation to characteristics other than gender. See also RRT Reference: N96/82294 (18 June 1997) and RRT Reference: N98/22513 (23 April 1999).
17. See, for instance, RRT Reference N98/21419 (11 January 1999) and RRT Reference N98/24944 (14 January 1999).
18. See, by analogy, the transcript of the hearing in MIMA v Abdi (1999) 87 FCR 280.
19. Immigration and Refugee Board of Canada, Guidelines Issues by the Chairperson Pursuant to s 65(3) Immigration Act, Guideline 4, Women Refugee Claimants Fearing Gender-Related persecution, 1997, at 5.
20. Chan v Canada [1995] 3 SCR 593.
21. See, for instance, Islam; Ex parte Shah per Lord Hope at 1038.
22. RRT Reference N93/000656 (Aug 3, 1994).
23. Lazo-Mojano v INS, 813 F.2d 1432 (9th Cir. 1987); Lopez Gularzo v INS, 99 F.3d 954 (9th Cir, 1996). See also Macklin, "Cross-border shopping for ideas: a critical review of United States, Canadian, and Australian approaches to gender-related asylum claims", at 40, 53-55.
24. Justice Sedley in Reg v Immigration Appeal Tribunal, Ex parte Shah [1997] Imm. A.R. 145 at 153 quoted in Islam; Ex parte Shah [1999] 2 WLR 1015 per Lord Steyn at 1028.
25. Ibid., per Lord Hoffman at 1032. See also Lord Steyn at 1021.
26. Ibid., per Lord Millett at 1041.
27. Ibid., per Lord Steyn at 1028.
28. Ibid., per Lord Hoffman at 1035.
29. Id.
30. Ibid., per Lord Millett at 1045.
31. Ram v MIEA & Anor (1995) 57 FCR per Burchett J at 568.  Approved in Applicant A v MIEA (1997) 190 CLR 225 at 284.
32. Islam; ex parte Shah per Lord Millett at 1042.
33. Id.
34. Harvey, C "Mainstreaming Gender in the Refugee protection process" New Law Journal 1999 April 9 at 534.