International

INTERNATIONAL ASSOCIATION OF REFUGEE LAW JUDGES OTTAWA CONFERENCE (OCTOBER 1998)

INTERIM REPORT ON MEMBERSHIP OF A PARTICULAR SOCIAL GROUP

PREPARED FOR THE INTER-CONFERENCE WORKING PARTY ON MEMBERSHIP OF A PARTICULAR SOCIAL GROUP BY:

Rodger Haines (Rapporteur) Deputy Chairperson Refugee Status Appeals Authority New Zealand October 1998 INDEX

PREFACE

INTRODUCTION

THE PRINCIPLES OF TREATY INTERPRETATION: THE VIENNA CONVENTION ON THE LAW OF TREATIES

THE PRINCIPLES OF INTERPRETATION: THE REFUGEE CONVENTION

EXTERNAL PERCEPTIONS OR INTERNALLY DEFINED CHARACTERISTICS

DEFINING THE GROUP BY THE PERSECUTION

VOLUNTARY ASSOCIATION

OVERLAPPING GROUNDS OF PERSECUTION

PREFACE

The Inter-conference Working Party on membership of a particular social group began its work in December 1997. Apart from the difficulty inherent in the subject matter, the Working Party has encountered other problems. First, the geographic spread of its membership is narrow. The whole of Continental Europe is unrepresented, apart from one Judge from the Netherlands. Second, there are no representatives from Canada, Australia, Africa, Central and South America. Third, as busy decision-makers, members have had but limited time to devote to what is a complex subject. Prohibitive costs have made it impracticable to meet.

In the result, this interim report represents not the considered views of the Working Party, but rather the inchoate first impressions of a difficult terrain yet to be surveyed.

Working Party members who have contributed to this interim report are (in alphabetical order):

Mr Richard Chalkey (UK)
Mr John Freeman (UK)
Mr Rodger Haines (NZ) - Rapporteur
Judge AHN Kruijer (Netherlands)
Mrs Rosy N Mannion (UK)
Ms Lory D Rosenberg (USA)
Sir Stephen Sedley (UK)
Ms Shahrzad Tadjbakhsh (UNHCR)

INTRODUCTION

"Courts and jurists have taken widely differing views as to what constitutes "membership of a particular social group" for the purposes of the Convention. This is not surprising. The phrase is indeterminate and lacks a detailed legislative history and debate. Not only is it impossible to define the phrase exhaustively, it is pointless to attempt to do so".

[1] So wrote McHugh J of the High Court of Australia in the course of delivering one of the majority judgments in Applicant A v Minister of Immigration and Ethnic Affairs (1997) 190 CLR 225, 259; (1997) 142 ALR 331, 355. In this case, the High Court of Australia split 3:2 on the issue whether a couple fearing sterilization under the one child policy of the People's Republic of China were refugees within the meaning of Article 1A(2) of the Refugee Convention. The majority (Dawson, McHugh & Gummow JJ), in separate judgments, held that the claim failed as it is not permissible to define a particular social group by reference to the act which gives rise to the well-founded fear of persecution.

[2] Ironically, the principles of treaty interpretation carefully delineated by McHugh J in his judgment at 251-256; 349-352 and which ultimately led to his dismissing the claim were expressly adopted by Brennan CJ at 231; 332 in the course of his dissenting judgment which held in favour of the claim. The second dissentient, Kirby J, while rejecting the expansive approach adopted by Brennan CJ, nevertheless also ruled in favour of the claim as he was of the view that the majority approach was too narrow (see 306-309; 393-395). Yet in two of the majority judgments (Dawson J 241; 340-341 & McHugh J 260; 355) there is express acceptance that the phrase "particular social group" should be given a broad interpretation. In short, the case affords a striking example of how a narrow issue (whether a social group can be defined by reference to the act that gives rise to the fear of persecution) can lead to widely diverging opinions premised on precisely the same interpretative principles.

[3] At one point, it was thought that in Canada (Attorney General) v Ward [1993] 2 SCR 689; (1993) 103 DLR (4^th) 1, the Supreme Court of Canada had achieved a workable framework for the interpretation and application of the social group category. La Forest J, writing for the Court, rejected the argument that this category is a safety-net intended as a catch-all for all the bases of persecution not included in the other four categories of race, religion, nationality and political opinion. The Supreme Court found that the Refugee Convention has built-in limitations, reflecting the fact that the international community did not intend to protect all suffering individuals (731; 28). The social group category is limited by the general boundary of the objectives of the Convention, namely the enjoyment of fundamental human rights and freedoms without discrimination (733-734; 29-30). The manner in which groups are distinguished for the purposes of discrimination law can be imported into this area of refugee law (735; 30). The Court formulated a working rule which identifies three possible categories of social groups (739; 33-34):

"1. Groups defined by an innate or unchangeable characteristic;

2. Groups whose members voluntarily associate for reasons so fundamental to their human dignity that they should not be forced to forsake the association; and

3. Groups associated by a former voluntary status, unalterable due to its historical permanence."

[4] Yet the application of these categories has proved controversial, even at the highest judicial level in Canada itself. Again, the challenge has been presented by one child policy claims. In Chan v Canada (Minister of Employment and Immigration) [1995] 3 SCR 593 the Supreme Court of Canada split 4:3. The majority, assuming without deciding that the claimant was a member of a particular social group, held that it had not been established that the fear of persecution was well-founded. They did not therefore address the issue on which the case had largely turned in the Federal Court of Appeal, namely whether a group defined solely by the fact that members face a particular form of persecution is a social group for the purpose of the Convention and whether it is a relevant distinction that a person is at risk not for what the person is, but for what that person has done. Heald and Desjardin JJA (Mahoney JA dissenting) in the Federal Court of Appeal had held that persecution could not solely define the group and a person at risk for what he or she has done (or does) is not a member of a particular social group.

[5] In the Supreme Court, La Forest J (who together with L'Heureux-Dubé & Gonthier JJ formed the dissenting minority), acknowledged (643) that in Ward he had himself drawn a distinction between what one does as opposed to what one fundamentally is, but emphasized that this simplified distinction was never intended to replace the Ward categories, which themselves were no more than a working rule and not to be applied in an unyielding, deterministic manner (642).

[6] In the very near future, the House of Lords will be in a position to make its own contribution to the social group jurisprudence. In the first week of November 1998, the House of Lords will hear the appeal from the judgment of the Court of Appeal reversing Sedley J in R v Immigration Appeal Tribunal, Ex parte Shah [1998] 1 WLR 74; (pet. all.) [1998] 1 WLR 270 (HL); [1997] Imm AR 145 (Sedley J), 584 (CA). One of the issues is whether the common uniting attribute must exist independently of the feared persecution. The putative social group is said to be Pakistanti women accused of transgressing social mores (adultery, disobedience to husbands) who are unprotected by their husbands or other male relatives.

[7] In the USA, as in Australia and Canada, divergence also appears to be the leitmotiv. In Lwin v INS 144 F. 3d 505 (7^th Cir. May 15, 1998) the Seventh Circuit has joined the First and Third Circuits in concluding that the Board of Immigration Appeals' definition of a particular social group as one that shares a "common immutable characteristic", (originating in Matter of Acosta 19 I&N Dec 211 (BIA 1985)) is central to the determination of what constitutes a particular social group. The Court rejected both the Ninth Circuit's voluntary associational relationship test⁽¹⁾, and the Second Circuit's test which emphasizes external perceptions of the existence of a group, by defining a particular social group as consisting of individuals who possess some fundamental characteristic in common that distinguishes them in the eyes of either the persecutor or the outside world.⁽²⁾ In embracing the approach of the First and Third Circuits, the Court stated that the common characteristic that defines the group must be one that the members of the group either cannot change, or should not be required to change because it is fundamental to their individual identities or consciences.⁽³⁾

[8] In some countries, notably Canada, the United States of America and Australia, the difficulties associated with the social group category have led to the publication of "Guidelines" or the enactment of domestic legislation or both. Similarly, in Europe there is the Joint Position on the Harmonized Application of the Term "Refugee" in Article 1 of the Geneva Convention of 28 July 1951 Relating to the Status of Refugees adopted by the Justice and Home Affairs Council of the European Union on 4 March 1996 on the basis of Article K.3 of the Treaty on European Union.⁽⁴⁾ Para 7.5 of the Joint Position addresses the social group category but in language so general it is of limited assistance. Indeed, much of the European jurisprudence is of limited assistance. For example, the view expressed by the Working Party member from the Netherlands concerning the jurisprudence of her country is that there are only a few decisions in which an explicit reference to particular social groups can be found; most of the relevant decisions having been given in the 1980s and one in 1996. Those decisions often concern homosexuals or have to do with the nationality of the spouse of the claimant. In Austria, the Administrative Court has always been very hesitant to make use of the social group category.⁽⁵⁾ In Belgium the Commission permanente de recours des réfugiés has adopted a formula very similar to that given by Professor James C Hathaway in The Law of Refugee Status (Butterworths 1991), namely that the notion of social group implies the presence of unifying factors between the members of a group like culture, linguistic origin, education, values and interests shared, as well as the possibility of identifying this group with regard to its relations with other social groups and with regard to the relations held with the authorities of a state. A social group can be a group characterised either by innate and unalterable characteristics, or by anterior characteristics that cannot be changed by members (ie, their history and former experiences). It may also be a voluntarily composed group provided that the purpose of the group is so fundamental to their human dignity that one cannot demand that it be renounced.⁽⁶⁾

[9] In Switzerland, the concept of the social group is based on the existence of common social characteristics that constitute a reason for illegal persecution.⁽⁷⁾ Of the jurisprudence in Germany, it has been said that the courts have not attempted to define a social group. Rather, the courts reach their decisions based on an intuitive sense of persecution based on membership in a social group.⁽⁸⁾ The same study points out that two German courts which have attempted to define the social group term have formulated two very different tests. A Hanover Court has focused on the internal structure of the putative group, where as a Court in Wiesbaden has focused on external perceptions, namely society's view of the group in question. In Denmark, the Refugee Appeals Board interprets the social group ground strictly. A social group is a group of people with a homogeneous background, behaviour and social status.⁽⁹⁾ In Spain, by contrast, no relevant decision based on membership of a particular social group can be found.⁽¹⁰⁾ In Italy, membership of a social group may simply be attributed to the victimized person or group by the persecutor. In some cases, the social group may not have existed previously but may be determined by the common characteristics of the victimized persons because the persecutor sees them as an obstacle to achieving his aims.⁽¹¹⁾

[10] Against this confused and changing jurisprudential backdrop, the Working Party has faced considerable challenges in analysing the current trends and developments and in identifying areas of divergence and areas of convergence. What follows is an interim report which attempts no more than an introduction to the issues.

[11] The interim report first considers the principles of treaty interpretation which apply to the Refugee Convention, and to what ends the courts have used those principles. It then looks at the distinction between external perceptions and internally defined characteristics, and the question of whether a group may legitimately be defined by the persecution it faces. The report also examines the courts' treatment of groups formed by voluntary association. Next, the relationship between membership of a particular social group and other Convention grounds is considered. Finally, the report assesses the role of guidelines and domestic legislation in influencing the courts' interpretation of membership of a particular social group.

[12] A framework for future work in this area is suggested in the Conclusion. If the discussion which follows appears to be somewhat focused on Canadian and Australian jurisprudence, that is because it is in these two countries that most of the cutting edge development has taken place.

THE PRINCIPLES OF TREATY INTERPRETATION: THE VIENNA CONVENTION ON THE LAW OF TREATIES

[13] Despite their relevance, Articles 31, 32 and 33 of the Vienna Convention on the Law of Treaties have received virtually no attention in the debate as to the meaning and application of the social group category. On one view, this is a surprising omission given that Article 31 is the leading general rule of interpretation of treaties and further given that the travaux préparatoires, so often referred to (but to little effect), is assigned only a supplementary role by Article 32.⁽¹²⁾

[14] The singular exception is Applicant A v Minister for Immigration and Ethnic Affairs (1997) 190 CLR 225; (1997) 142 ALR 331 (HCA) where McHugh J (one of the majority) at 251-256; 349-352 expressly addressed the Vienna Convention in the context of not only the Refugee Convention, but also in the context of a treaty provision enacted in or as part of a domestic statute. Brennan CJ, while dissenting as to the result, concurred with this aspect of the judgment of McHugh J (231; 332) as did Gummow J (one of the majority) at 277; 370. Dawson J (also one of the majority) separately addressed the issue at 240-241; 339-340 as did Kirby J (dissenting) at 294; 383-384.

[15] McHugh J at 252-253; 350-351 noted the three separate but related principles contained in Article 31. First, an interpretation must be in good faith, which flows directly from the rule pacta sunt servanda. Secondly, the ordinary meaning of the words of the treaty are presumed to be the authentic representation of the parties' intentions. This principle has been described as the "very essence" of a textual approach to treaty interpretation. Thirdly, the ordinary meaning of the words is not to be determined in a vacuum removed from the context of the treaty or its object or purpose. After referring to the controversy whether textual interpretation takes precedence over the object and purpose of the treaty, McHugh J preferred the ordered yet holistic approach taken by Zekia J in Golder v United Kingdom (1975) 1 EHRR 524, 544 (ECHR). That is, primacy is to be given to the written text of the Convention, but the context, object and purpose of the treaty must also be considered.

[16] While it might be observed that agreement on the principles of treaty interpretation conspicuously failed to assist the five judges of the High Court of Australia in reaching agreement on the meaning and application of the social group category, the approach is nevertheless a principled one and it avoids the dangers inherent in resorting (inappropriately) to domestic legislation, as arguably occurred in Canada (Attorney General) v Ward [1993] 2 SCR 689, 738; (1993) 103 DLR (4th) 1, 32.

[17] The ordered yet holistic approach preferred in Applicant A has been adopted by the New Zealand Refugee Status Appeals Authority in Refugee Appeal No. 70366/96 Re C (22 September 1997).

[18] What conclusions can be drawn from the text, object and purpose of the Refugee Convention?

THE PRINCIPLES OF INTERPRETATION: THE REFUGEE CONVENTION

[19] In both Ward and Applicant A the conclusion reached was that the category of "particular social group" was added with the intention of broadening the reach of the other four grounds of race, religion, nationality and political opinion.

[20] See Ward 730; 27 where it was stated that:

"The fact that this class was added to enlarge the range of cases falling within the definition of 'refugee' therein was initially a Cold War reaction aimed at ensuring a haven for capitalists fleeing the persecution they encountered in Eastern Bloc regimes after World War II ... The scope of 'particular social group', however, was not meant to be limited to that specific historical circumstance and no-one has ever so contended. The ambit of this portion of the definition ... must be evaluated on the basis of the basic principles underlying the treaty."

[21] To similar effect see the majority judgments in Applicant A, and in particular, Dawson J at 241; 340 and McHugh J at 259, 265-266; 355, 360.

[22] While both the Supreme Court of Canada in Ward and the majority of the High Court of Australia in Applicant A saw the Refugee Convention as necessarily limited in the protection it affords to refugees, the two courts nevertheless adopted different paths to this conclusion.

[23] The Ward approach (731-733; 28-29) gives express recognition to the limitations built into the protection regime afforded by the Refugee Convention. For instance, the need for "persecution" excludes economic migrants and victims of natural disasters. Secondly, the enumeration of only five grounds of persecution further limits the responsibility of State Parties. In addition, the Preamble to the Convention outlines the boundaries of the objectives sought to be achieved by the Convention, namely the assurance of basic human rights without discrimination (733; 29). The Supreme Court of Canada then built on Matter of Acosta 19 I&N Dec 211 (BIA 1985), the decision of the US Board of Immigration Appeals which identifies the grounds of race, religion, nationality and political opinion as being characteristics which one either cannot change or ought not to be required to change. Applying the ejusdem generis rule of interpretation, the Board found that the five Convention grounds are all defined by a characteristic which is unchangeable or from which disassociation cannot be expected. In the result, a characteristic which can be changed, or from which disassociation is possible is not included in the Refugee Convention (737-738; 32-33):

"Canada's obligation to offer a haven to those fleeing their homelands is not unlimited. Foreign Governments should be accorded leeway in their definition of what constitutes anti-social behaviour of their nationals. Canada should not overstep its role in the international sphere by having its responsibility engaged whenever any group is targeted. Surely there are some groups, the affiliation in which is not so important to the individual that it would be more appropriate to have the person dissociate him or herself from it before Canada's responsibility should be engaged. Perhaps the most simplified way to draw the distinction is by opposing that which one is against that which one does, at a particular time. For example, one could consider the facts in Matter of Acosta, in which the claimant was targeted because he was a member of a taxi driver co-operative. Assuming no issues of political opinion or the right to earn some basic living are involved, the claimant was targeted for what he was doing and not for what he was in an immutable or fundamental way." [emphasis in original]

[24] While the majority in Applicant A do not directly challenge the approach taken in Ward, they do question aspects of it.

[25] In particular, there was a challenge to whether a genus can be found in the phrase "race, religion, nationality, membership or a particular social group or political opinion". See McHugh J at 263; 358. This point appears to have been overlooked in Secretary of State for the Home Department v Savchenkov [1996] Imm AR 28 (CA). Second, there was by no means agreement that a valid distinction can be drawn between what a person is and what a person has done or does. See Dawson J at 242-243; 342, a view shared by La Forest J (dissenting) in Chan v Canada (Minister of Employment and Immigration) [1995] 3 SCR 593, 643-644 (SC:Can). But thirdly, and most importantly, the majority in Applicant A was of the clear view that the characteristic or element which unites the group cannot be a common fear of persecution. Dawson J at 242; 341 stated:

"There is more than a hint of circularity in the view that a number of persons may be held to fear persecution by reason of membership of a particular social group where what is said to unite those persons into a particular social group is their common fear of persecution. A group thus defined does not have anything in common save fear of persecution, and allowing such a group to constitute a particular social group for the purposes of the Convention 'completely reverses the statutory definition of Convention refugee in issue (wherein persecution must be driven by one of the enumerated grounds and not vice versa).'"⁽¹³⁾

And at 243; 342:

"In this case, the reason the appellants fear of persecution is not that they belong to any group since there is no evidence that being the parents of one child and not accepting the limitations imposed by government policy is a characteristic which because it is shared with others, unites a collection of persons and sets them apart from society at large. It is not an accurate response to say that the government itself perceives such persons to be a group and persecutes individuals because they belong to it. Rather, the persecution is carried out in the enforcement of a policy which applies generally. The persecution feared by the appellants is a result of the fact that, by their actions, they have brought themselves within its terms. The only recognizable group to which they can sensibly be said to belong is the group comprising those who fear persecution pursuant to the one child policy. For the reasons I have given, that cannot be regarded as a particular social group for the purposes of the Convention."

[26] As to the anti-discrimination notions in the Preamble relied upon by the Supreme Court of Canada in Ward, Dawson J and Gummow J (members of the majority) at 248; 346 and 283; 374 relied on the very same paragraph, but without emphasizing the phrase "without discrimination". The relevant paragraph in the Preamble reads:

"Considering that the Charter of the United Nations and the Universal Declaration of Human Rights ... have affirmed the principle that human beings shall enjoy fundamental rights and freedoms without discrimination."

Dawson J at 248; 346 stated

"However, what I have said is sufficient to illustrate the simple point that despite the reference in the Convention to the concern that persons enjoy the 'widest possible exercise of ... fundamental rights and freedoms', there are limits on the extent to which the Convention attempts to translate that concern into practical reality. In that respect, the Convention, like many international and municipal instruments, does not necessarily pursue its primary purpose at all costs. The purpose of an instrument may instead be pursued in a limited way, reflecting the accommodation of differing view points, the desire for limited achievement of objectives, or the constraints imposed by limited resources. No doubt many of those limits in the present context spring from the well-accepted fact that international refugee law was meant to serve as a 'substitute' for national protection where the latter was not provided due to discrimination against persons on grounds of their civil and political status. It would therefore be wrong to depart from the demands of language and context by invoking the humanitarian objectives of the Convention without appreciating the limits which the Convention itself places on the achievement of them."

[27] The underlying rationale was explained by McHugh J at 263; 358 as being that allowing persecutory conduct of itself to define a particular social group would, in substance, permit the "particular social group" ground to take on the character of a safety-net. It would impermissibly weaken, if not destroy, the cumulative requirements of "fear of persecution", "for reasons of" and "membership of a particular social group" in the definition of "refugee". It would also effectively make the other four grounds of persecution superfluous. It would mean that persons who had a well-founded fear of persecution were members of a particular social group because they feared persecution. But the textual approach to interpretation dictated that the only persecution that is relevant is persecution for reasons of membership of a group which means that the group must exist independently of, and not be defined by the persecution.

[28] Developing this point at 264; 359, McHugh J arrived at the heart of his analysis:

"Only in the 'particular social group' category is the notion of 'membership' expressly mentioned. The use of that term in conjunction with 'particular social group' connotes persons who are defined as a distinct social group by reason of some characteristic, attribute, activity, belief, interest or goal that unites them. If the group is perceived by people in the relevant country as a particular social group, it will usually but not always be the case that they are members of such a group. Without some form of internal linking or unity of characteristics, attributes, activities, beliefs, interests or goals, however, it is unlikely that a collection of individuals will or can be perceived as being a particular social group."

[29] From this brief analysis it can be seen that the Canadian and Australian approaches differ. Whereas the boundaries to the Refugee Convention, and in turn the social group category, were found by the Supreme Court of Canada in Ward to lie in the limitations inherent in the concept of persecution, the anti-discrimination notions underlying the Refugee Convention and in the ejusdem generis rule, the majority of the High Court of Australia in Applicant A gave primary emphasis to the text, object and purpose of the Convention. They shifted the emphasis away from anti-discrimination notions and disputed the application of the ejusdem generis rule.

[30] But on one point both Courts agreed (with only Brennan CJ at 235-236; 336 dissenting), namely that the so-called safety-net interpretative approach was impermissible. See Ward 731; 28:

"Although the delegates inserted the social group category in order to cover any possible lacuna left by the other four groups, this does not necessarily lead to the conclusion that any association bound by some common threat is included. If this were the case, the enumeration of these bases would have been superfluous; the definition of 'refugee' could have been limited to individuals who have a well-founded fear of persecution without more. The drafters' decision to list these bases was intended to function as another built-in limitation to the obligations of signatory states. The issue that arises, therefore, is the demarcation of this limit."

[31] The significant point made sub silentio in Applicant A, however, is that the rejection of the safety-net approach necessarily implies that a social group cannot be formulated in terms which have the effect of re-introducing the previously rejected safety-net approach. Thus persecution alone cannot define the group. See Dawson J at 242; 341-342:

"... if a shared fear of persecution were sufficient to constitute a particular social group, it would render at least three of the other four Convention reasons - race, religion and nationality - superfluous. It is one thing to say that the five Convention reasons can overlap; it is quite another to construe one of them in a manner which renders three of the others unnecessary and the fourth - political opinion - almost so. To construe the term 'particular social group' in that way would make it an almost all-encompassing safety-net, allowing a persecutory law or practice of general application to constitute those whose actions bring themselves within its terms members of a particular social group. Such a construction would be contrary to the context in which the words 'particular social group' appear."

[32] The majority could not, for this reason, agree with the dissenting judgment of La Forest J in Chan. Indeed, La Forest J had earlier in Ward at 729; 26 expressly recognized that people cannot be characterized as a particular social group merely by virtue of their common victimization as the objects of persecution.

[33] The troublesome nature of this category is further exemplified by the fact that while Brennan CJ (dissenting) expressly embraced the safety-net approach (235-236; 336), Kirby J, the other dissentient, rejected it (294; 383). Yet Kirby J (307; 393) viewed favourably the possibly less-principled "case by case" or intuitive approach to the problem which is suggested by the German courts:

"In the result, I agree with the opinion of Professor Fullterton citing the conclusion of Graves. Some of the groups to which the definition applies are voluntary; others are not. Some are cohesive; others are not. Some are homogeneous; others are not. Some involve immutable characteristics; others do not. Some involve characteristics central to the members' identities; others do not. Professor Graves urges that courts and agencies should turn away from attempts to formulate abstract definitions. Instead, they should recognize 'particular social groups' on a case by case basis. This approach conforms to the refusal of the German courts to attempt a definition, or exhaustive description of the category of 'particular social groups'. It accepts that an element of intuition on the part of decision-makers is inescapable, based on the assumption that they will recognize persecuted social groups of particularity when they seen them. While this is not an entirely satisfactory conclusion, it is preferable to an attempt by courts unduly to narrow the operation of the Convention or to impose upon its deliberately broad and ambulatory language categories which are by no means exhaustive of the actual words used. The development and expression of such categories, at least in the first instance, is the province of administrators and review tribunals with experience of refugee claims. It is not the task of appellate courts to whom these cases are but occasional visitors".

[34] While McHugh J himself at 259; 355 acknowledged the impossibility and pointlessness of attempting to define the phrase exhaustively, the difference between the majority and minority in Applicant A is not one of degree, but one of principle. Neither the text of the Convention nor its objects and purposes can be surrendered to an "intuitive" approach or to a complete re-writing of the text of the definition to make the social group category embracing of all circumstances of persecution. The solution may be difficult, but must nevertheless be attempted on a principled approach.

[35] In the result, the majority in Applicant A did not themselves set out to exhaustively define "membership of a particular social group". They possibly identified what it is not. The guide provided by Dawson J at 241; 341 is supported by similar comments by McHugh J at 266; 361 and Gummow J at 284-285; 375-376:

"A particular social group, therefore is a collection of persons who share a certain characteristic or element which unites them and enables them to be set apart from society at large. That is to say, not only must such persons exhibit some common element; the element must unite them, making those who share it a cognisable group within their society."

[36] But whereas Dawson J saw no reason to confine a particular social group to small groups or to large ones, McHugh J at 266; 360 thought that it was probably intended to cover only a relatively large group.

EXTERNAL PERCEPTIONS OR INTERNALLY DEFINED CHARACTERISTICS

[37] In Applicant A, the majority confirmed that members of a social group must have associative qualities that go to the members' identity. In other words, the association must go to what a person is rather than to what she or he does. In the words of McHugh J at 264; 359, without some form of internal linking or unity of characteristics, attributes, activities, beliefs, interests or goals it is unlikely that a collection of individuals will or can be perceived as being a particular social group. In New Zealand, the Refugee Status Appeals Authority in Refugee Appeal No. 1312/93 Re GJ (30 August 1995) 57 has similarly stated that there must be an internal defining characteristic shared by members of the particular social group.

[38] The concern appears to be that if the defining characteristic of a social group is the societal perception that the group exists, then almost everyone in a society could be said to be a member of a social group, if not several groups. Thus, the difficulty with the external perception approach is that it potentially enlarges the social group category to an almost meaningless degree. That is, by making societal attitudes determinative of the existence of the social group, virtually any group of persons in a society perceived as a group could be said to be a particular social group. If the safety-net approach is to be rejected, such rejection may also question the external perception approach on the grounds that it reintroduces the previously rejected safety-net approach. In this context the United States Court of Appeals, Second Circuit decision in Gomez v Immigration and Naturalization Service 947 F. 2d 660 (2d Cir. 1991) affords an example of a case in which the external perceptions of the existence of the group has been emphasized.

[39] At the end of the day, it may be a matter of emphasis and degree. For there is force in the observation made by McHugh J in Applicant A (264; 359) that while persecutory conduct cannot define the social group, the actions of the persecutors may serve to identify or even cause the creation of a particular social group in society:

"The fact that the actions of the persecutors can serve to identify or even create 'a particular social group' emphasizes the point that the existence of such a group depends in most, perhaps all, cases on external perceptions of the group. The notion of persecution for reasons of membership of a particular social group implies that the group must be identifiable as a social unit. Only in the 'particular social group' category is the notion of 'membership' expressly mentioned. The use of that term in conjunction with 'particular social group' connotes persons who are defined as a distinct social group by reason of some characteristic, attribute, activity, belief, interest or goal that unites them. If the group is perceived by people in the relevant country as a particular social group, it will usually but not always be the case that they are members of such a group. Without some form of internal linking or unity of characteristics, attributes, activities, beliefs, interests or goals, however, it is unlikely that a collection of individuals will or can be perceived as being a particular social group. Those indiscriminately killed or robbed by guerillas, for example, are not a particular social group.

[40] The group may qualify as a particular social group, however, even though the distinguishing features of the group do not have a public face. It is sufficient that the public is aware of the characteristics or attributes that, for the purposes of the Convention, unite and identify the group." [emphasis in the original]

[41] In addition to these factors, McHugh J pointed out (265; 360) that it is possible to have an imputed social group. That is, it is not necessary that the group should possess the attributes that they are perceived to have.

[42] The point has been more recently illustrated by Minister for Immigration and Multicultural Affairs v Zamora [1998] 913 FCA (5 August 1998) in which Black CJ, Branson & Finkelstein JJ of the Federal Court of Australia addressed the question whether tourist industry workers in Ecuador were a particular social group because they were targeted by criminal gangs seeking to facilitate the robbing of tourists. The Court, while accepting that the group had a distinguishing attribute, added that it was doubtful that such a group would be one recognizable in Ecuadorian society as one whose members share something which unites them:

"There will no doubt be cases in which persons who have in common no more than a shared occupation do form a cognisable group in their society. This may well come about, as McHugh J recognized in Applicant A's case, when persons who follow a particular occupation are persecuted by reason of the occupation that they follow. The persecution for following a particular occupation may well create a public perception that those who follow the occupation are a particular social group. Human rights workers in certain nations subject to totalitarian rule come to mind as a possible example. Ordinarily however, persons who have in common no more than a shared occupation are not recognizable as a particular social group in their society. That is, they are not defined as individuals in any meaningful way by reason of their occupation. In the words of Gummow J in Applicant A's case, they are simply a 'broadly defined segment of those sharing a particular country of nationality'." [emphasis in original].

[43] What is possibly seen in this context is an attempt to find a middle way between the Scylla of over-inclusion and the Charybdis of irrelevance.

DEFINING THE GROUP BY THE PERSECUTION

[44] Dawson, McHugh and Gummow JJ in Applicant A emphatically rejected the argument that persecutory conduct itself could define a particular social group. The same approach has been taken in New Zealand by the Refugee Status Appeals Authority in Refugee Appeal No. 3/91 Re ZWD (20 October 1992) 83-84 in a case involving the one child policy. It is also the approach taken to date in the United Kingdom. See Secretary of State for the Home Department v Savchenkov [1996] Imm AR 28, 37, 38 (CA) (security guard approached by the Mafia to act as an informer) and R v Immigration Appeal Tribunal, Ex parte Syeda Shah [1998] 1 WLR 74, 89 (CA) (Pakistani women accused of transgressing social mores). The House of Lords will next month have to grapple with the latter issue.

[45] French jurisprudence would appear to stand in stark contrast to the approach taken by the High Court of Australia and the English Court of Appeal. In a case involving an Algerian transsexual (Ourbih) the Commission de recours des réfugiés (CRR) at first rejected the case in 1995 by concluding that the marginalisation of the applicant in Algerian society because he was a transsexual was not enough for him to viewed as being a member of a particular social group. When the case reached the Conseil d'Etat, the view expressed by the Commissaire du Gouvernment: J D Combrexelle, 23 June 1997 was that a social group in the context of the Refugee Convention does not detach itself from the persecution to which it is an object and that one could even say that it is the persecution which is the constitutive element in determining the social group. On such analysis, transsexuals would only constitute a particular social group in countries where their sexual orientation leads to persecution. The Conseil d'Etat agreed with this analysis and requested the CRR to re-examine the case because transsexuals, whose common characteristics may set them apart from others in society, could constitute a particular social group, but only if they face persecution. Subsequently, the CRR, upon reconsidering the case, came to the conclusion that transsexuals, because of their common characteristics different from others, are in fact exposed to persecution in Algeria, and as such they constitute, under such conditions, a particular social group.⁽¹⁴⁾

[46] It would appear that in Italy also persecutory conduct can itself define a particular social group.⁽¹⁵⁾

VOLUNTARY ASSOCIATION

[47] Two of the three Ward categories refer to an existing or former voluntary status. It would appear that they were framed around the premise (738; 33) that there are some groups, the affiliation in which is not so important to the individual that it would be more appropriate to have the person disassociate him or herself from it before engaging the responsibility of the State Parties to the Convention. Thus, Ward himself could have left the violent terrorist group to which he belonged without abdicating his human dignity. Similarly the taxi driver in Matter of Acosta could avoid risk of harm by leaving the taxi co-operative.

[48] Dawson J, one of the majority in Applicant A at 241; 341, was of the view that there was nothing to suggest that the uniting particular be voluntary and in this regard he and McHugh J at 260; 356 rejected as far too narrow the "voluntary association or relationship" test of Sanchez-Trujillo v Immigration and Naturalization Service 810 F. 2d 1571 (9^th Cir. 1986). In Savchenkov at 36 one of the reasons against finding that persons threatened by the Mafia were a social group was that there was no evidence that the persons so threatened ever associated with one another.⁽¹⁶⁾

[49] This may perhaps be missing the point. There are two distinct issues. First, whether in law and in fact a particular social group exists (and that the refugee claimant is a member of it). Second, whether by leaving the group the claimant can eliminate the risk of persecution. It is in the context of this second issue that the question of abdicating human dignity is relevant.

[50] Since Matter of Acosta this appears to have been the general approach of the United States Board of Immigration Appeals and of the US Courts. Nevertheless, the United States Court of Appeals for the Ninth Circuit in Li v Immigration and Naturalization Service 92 F. 3d 985 (9^th Cir. 1996) rejected a social group claim made by an individual who claimed that he was a member of a social group comprising Chinese citizens with low economic status. Rather than address whether that characteristic was changeable or capable of change, the Court found that the group would not be recognized because the association was involuntary. That seems to contrast directly with the concept of the characteristic being immutable or it being unconscionable to expect the individual to change.

OVERLAPPING GROUNDS OF PERSECUTION

[51] It is often the case that the Convention grounds will overlap. The UNHCR Handbook on Procedures and Criteria for Determining Refugee Status (1979) para 78 explains that persecution based on membership of a particular social group frequently occurs because the government lacks confidence in the group's loyalty or perceives the group's political outlook or economic activity as an obstacle to government policies. That is, the Handbook explicitly recognizes that in some circumstances there is a distinct political ingredient to the social group term. The significance of this point is highlighted in the gender persecution context. One of the often repeated complaints is that too often in refugee determination, women's experiences are not recognized as "political" due to a skewed interpretation of a largely male orientated body of law which privileges male-dominated public activities over the activities of women which take place largely in the private sphere. Commentators complain that in some countries there has been a failure to recognize the political nature of seemingly private acts of, and harm to, women.

[52] The New Zealand Refugee Status Appeals Authority attempted to address some of these concerns in Refugee Appeal No. 2039/93 Re MN (12 February 1996) 49-52. In doing so, it accepted two points made by Professor Maryellen Fullerton in "A Comparative Look at Refugee Status Based on Persecution Due to Membership in a Particular Social Group" 26 Cornell Int'l LJ 505, 550 & 552 (1993). First, it is just as important to prevent a claim of persecution based on political opinion from masquerading as persecution based on social group, as it is to prevent a social group from masquerading as a political opinion case. Second, where persecution focused on members of a social group has an implicit political dimension, proving that the group is persecuted for its actual or imputed political opinion might be difficult. The concept of a social group is therefore a needed one in that it fills a noticeable gap in the categories of victims of persecution.

GUIDELINES

[53] In some jurisdictions, notably Canada,⁽¹⁷⁾ United States of America⁽¹⁸⁾ and Australia,⁽¹⁹⁾ guidelines have been issued with specific reference to women refugee claimants. These guidelines can have a measurable impact on the interpretation and application of the social group category.⁽²⁰⁾ In this regard, contrast the direct references to the Canadian and US guidelines in the concurring opinion of Lory D Rosenberg in Matter of Kasinga Interim Decision 3278 (13 June 1996)⁽²¹⁾ with the more circumscribed views to be found in the concurring opinion of Lauri Steven Filppu (joined by Michael J Heilman).

[54] Guidelines have also been issued in respect of children by the UNHCR,⁽²²⁾ Canada⁽²³⁾ and the argument in favour of the USA following suit is made by Jacqueline Bhabha & Wendy A Young, "Through a Child's Eyes: Protecting the Most Vulnerable Asylum-Seekers" 75 Interpreter Releases 757 (June 1, 1998).

[55] The European Union Joint Position is, on one view, a guideline of a kind, but not all would necessarily accept without reservation the statement by the English Court of Appeal in R v Secretary of State for the Home Department, Ex parte Robinson [1997] 3 WLR 1162, 1169; [1997] 4 All ER 210, 216 (CA) that the Joint Position reflects a contemporary understanding of the obligations created by the Refugee Convention.

[56] For common law countries, the introduction of guidelines represents a substantial departure from the incremental development of the law through the hierarchy of tribunals and courts on a case by case basis. The long term impact of guidelines on this process has not yet been studied, nor does there appear to have been any vigorous critique of the process by which jurisprudential and "value" judgments are made in the process of setting the guidelines,⁽²⁴⁾ or having been made, how they are to evolve and change. Furthermore, if the underlying judicial system has so lost its way as to require the introduction of remedial guidelines, what does this say of the ability of that same system to adequately address other refugee issues? And by whose judgment is it to be said that the system has lost its way and cannot adequately address refugee issues without the use of guidelines? Unfortunately, many of these questions have yet to be addressed in the debate notwithstanding their importance.

[57] Guidelines are not, however, the only extra-judicial means of influencing the development of refugee law. Of equal, if not greater importance is the growing phenomenon of the enactment of domestic legislation which expands or contracts the Convention definition of "refugee". This process too can allow the intrusion of overtly political considerations into what should be an objective decision-making process.

LEGISLATION

[58] In Ireland, the Refugee Act 1996 s 1(1) specifically defines the phrase "membership of a particular social group" as to include gender and sexual orientation:

"'membership of a particular social group' includes membership of a trade union and also includes membership of a group of persons whose defining characteristic is their belonging to the female or the male sex or having a particular sexual orientation."

[59] In the United States of America, the Illegal Immigration Reform and Immigrant Responsibility Act of 1996 amended the definition of "refugee" contained in the Immigration and Nationality Act 1952. Specifically, § 601(a) of the 1996 Act added the following sentence to the end of INA § 101(a)(42)'s definition of refugee:

"For purposes of determinations under this Act, a person who has been forced to abort a pregnancy or to undergo involuntary sterilization, or who has been persecuted for failure or refusal to undergo such a procedure or for other resistence to a coercive population control programme, shall be deemed to have been persecuted on account of political opinion, and a person who has a well founded fear that he or she will be forced to undergo such a procedure or subject to persecution for such failure, refusal or resistence shall be deemed to have a well founded of persecution on account of political opinion."

[60] The section goes on to provide, however, that there is a numerical cap of 1,000 refugees who may be granted asylum under this provision. Section 601(b) states:

"For any fiscal year, not more than a total of 1,000 refugees may be admitted under this sub-section or granted asylum under s 208 pursuant to a determination under the third sentence of section 101(a)(42) (relating to persecution for resistence to coercive population control methods)."

[61] This provision, which took effect immediately upon enactment, effectively overrules the Board of Immigration Appeals precedent decisions on coercive family planning, including Matter of Chang (BIA 1989) and Matter of G- (BIA 1993). In those cases, the Board had held that implementation of population control policies by coercive means is not on its face a basis for asylum eligibility, even to the extent that involuntary sterilizations may occur.⁽²⁵⁾

[62] In Australia, following the delivery of judgment by Sackville J at first instance in Applicant A,⁽²⁶⁾ the Minister not only appealed, he also introduced the Migration Legislation Amendment Bill (No.3) 1995 (Cth). The Bill would have amended the Migration Act to provide that the fertility control policies of foreign governments cannot be used to found a claim that a person belongs to a particular social group for the purposes of making out a claim for refugee status. In the event, the Bill was never put to a vote in Parliament because the Full Federal Court reversed the finding of Sackville J in a ruling subsequently confirmed by the High Court.⁽²⁷⁾

GENERAL CONCLUSIONS

[63] The social group ground is clearly a troublesome category and the problems are exacerbated by the fact that pressing contemporary issues are necessarily drawn to it, in particular, birth control policies, sexual orientation, gender persecution, violation of social mores, domestic violence and female genital mutilation.

[64] There is a tendency to look for a definition of "membership of a particular social group", but as Professor Guy S Goodwin-Gill cautions in The Refugee in International Law (2^nd ed, Oxford, Clarendon Press, 1996) at 365-366:

"Decision-makers may need to accept that there is probably no single coherent definition, but rather a set of variables, a 'range of permissible descriptors'."

[65] Much the same was said by McHugh J in Applicant A in the passage cited at the beginning of this interim report.

[66] The future direction of the endeavours of the Working Party may therefore lie not in setting out to define that which is itself already a definition, but in focusing on the characteristics which make up a particular social group, or identifying a range of factors which can be looked to in order to decide whether, on the particular facts, the refugee claimant is indeed a member of a particular social group.

[67] Not to be overlooked in this study is the significance of the terms "particular" and "social". But beyond this is the importance of the membership element (see McHugh J in Applicant A at 264; 359). Thereafter there is the interpretation of the all important phrase "for reason of" which requires a link to be established between the anticipated harm and the membership of the social group. But underlying all of these interpretive aspects is the broader conceptual issue inherent in the refugee definition itself, namely the nature and scope of state protection.

[68] Hopefully this interim report will provide a basis for a more detailed study in which these and many other related issues can be taken up and explored.

Rodger Haines (Rapporteur) Deputy Chairperson Refugee Status Appeals Authority New Zealand October 1998

1. Sanchez-Trujillo v Immigration and Naturalization Service 801 F. 2d 1571 (9^th Cir. 1986).
2. Gomez v Immigration and Naturalization Service 947 F. 2d 660 (2^d Cir. 1991).
3. See also 75 Interpreter Releases 1189 (Aug 31, 1998)
4. For the text of the Joint Position see Richard Plender, Basic Documents on International Law (2^nd rev ed, Martinus Nijhoff, Dordrecht 1997) 519.
5. Claus Hullmann, "Austria" in Carlier et al (eds) Who is a Refugee? A Comparative Case Law Study (Kluwer Law International 1997) 11, 49.
6. Dirk Vanheule, "Belgium" in Carlier et al (eds) Who is a Refugee? A Comparative Case Law Study (Kluwer Law International 1997) 57, 100. Contrast the statement made by the Belgium representative at the UNHCR Symposium on Gender-Based Persecution, Geneva, 22-23 February 1996 that in Belgium, systematic jurisprudence regarding gender-based persecution does not exist: (1997) IJRL Special Issue 40-41.
7. Klaus Hullman, "Switzerland" in Carlier et al (eds) Who is a Refugee? A Comparative Case Law Study (Kluwer Law International 1997) 111, 156.
8. Maryellen Fullerton, "A Comparative Look at Refugee Status Based on Persecution Due to Membership in a Particular Social Group" 26 Cornell Int'l LJ (1993) 505, 533. Cf Kaus Hullman, "Germany" in Carlier et al (eds) Who is a Refugee? A Comparative Case Law Study (Kluwer Law International 1997) 225, 283-284.
9. Pia Lynggaard Justesen, "Denmark" in Carlier et al (eds) Who is a Refugee? A Comparative Case Law Study (Kluwer Law International 1997) 291, 330.
10. Carlos Peña Galiano, "Spain" in Carlier et al (eds) Who is a Refugee? A Comparative Case Law Study (Kluwer Law International 1997) 337, 368.
11. (1997) IJRL Special Issue 57 (presentation by the Italian representative at the UNHCR Symposium on Gender-Based Persecution, Geneva, 22-23 February 1996).
12. See further Professor Guy S Goodwin-Gill, The Refugee in International Law (2^nd ed, Clarendon Press, Oxford 1996) 366-368.
13. The phrase in quotation marks has been taken from Chan v Canada [1993] 3 FC 675, 692-693 per Heald JA (FCA).
14. For a more detailed analysis of the French jurisprudence prepared by Working Party member, Shahrzad Tadjbakhsh, see the study attached to this paper as Appendix 1.
15. (1997) IJRL Special Issue 57 (presentation by the Italian representative at the UNHCR Symposium on Gender-Based Persecution, Geneva, 22-23 February 1996).
16. Contrast R v Immigration Appeal Tribunal, Ex parte Shah [1998] 1 WLR 74, 91 (CA) where Henry LJ appears to have accepted that non-cohesive, disparate individuals may, in certain situations, be capable of being a particular social group.
17. Women Refugee Claimants Fearing Gender-Related Persecution, March 9 1993 (1993) 5 IJRL 278. The guidelines are to be read with the Update published November 25 1996.
18. Immigration and Naturalization Service Gender Guidelines - Considerations for Asylum Officers Adjudicating Asylum Claims from Women (1995) 7 IJRL 700. The guidelines are also reproduced in 72 Interpreter Releases 781 (June 5, 1995).
19. Department of Immigration and Multicultural Affairs: Refugee and Humanitarian Visa Applicants - Guidelines on Gender Issues for Decision-Makers (July 1996) (1997) IJRL Special Issue 195.
20. See for example the discussion by Valerie L Oosterveld, "The Canadian Guidelines on Gender-Related Persecution: An Evaluation" (1996) 8 IJRL 569 and Deborah Anker, Nancy Kelly & John Willshire-Carrera, "The BIA's New Asylum Jurisprudence and its Relevance for Women's Claims" 73 Interpreter Releases 1173 (Sept 9, 1996).
21. Now reported at (1997) IJRL Special Issue 213, 233.
22. UNHCR, Refugee Children: Guidelines on Protection and Care (1994).
23. Child Refugee Claimants: Procedural and Evidential Issues (Guidelines Issued by the Chairperson Pursuant to s 65(3) of the Immigration Act: Guideline 3). In effect from September 30, 1996. See further Refuge Vol 15, No5 p 23 (1996).
24. Cf Deborah E Anker, "Women Refugees: Forgotten no Longer?" in Carlier & Vanheule (eds), Europe and Refugees: A Challenge? L'Europe et les réfugiés: Un défi? (Kluwer Law International 1997) 125, 134-135.
25. See further 73 Interpreter Releases 1582 (Nov. 11, 1996). As to the effect of this amendment, see further "BIA Decides Coercive Family Planning Case, Holds 1996 Act Supercedes Chang" 74 Interpreter Releases 115 (Jan. 13, 1997); "BIA Decides Coercive Family Planning, § 212(c) Cases" 74 Interpreter Releases 1073 (July 14, 1997); Paul Wickham Schmidt & Carolyn Anne Elliot, "Update on Precedent Decisions of the Board of Immigration Appeals" 74 Interpreter Releases 1665 (Nov. 3, 1997).
26. Minister for Immigration and Ethnic Affairs v Respondent A (1994) 127 ALR 383.
27. Mary Crock, "Apart from Us or a Part of Us? Immigrants' Rights, Public Opinion and the Rule of Law" (1998) 10 IJRL 49, 59 and see also Penelope Matthew, "Retreating from the Refugee Convention" in Alston & Chiam eds, Treaty-Making and Australia: Globalisation versus Sovereignty? (Federation Press, 1995) 149, 154-160.