Domestic violence and gender-based persecution: how refugee adjudicators judge women seeking refuge from spousal violence--and why reform is needed.
Subject: Tenants (Surveys)
Tenants (Crimes against)
Tenants (Political aspects)
Tenants (Political activity)
Tenants (Training)
Refugees (Crimes against)
Refugees (Surveys)
Refugees (Political aspects)
Refugees (Political activity)
Refugees (Training)
Abused women (Crimes against)
Abused women (Surveys)
Abused women (Political aspects)
Abused women (Political activity)
Abused women (Training)
Immigration policy (Political aspects)
Instrument industry (Political aspects)
Instrument industry (Political activity)
Instrument industry (Training)
Social norms (Political aspects)
Women's rights (Political aspects)
Emigration and immigration law (Political aspects)
Family violence (Political aspects)
International cooperation (Political aspects)
Emigration and immigration (Political aspects)
Canadians (Crimes against)
Canadians (Surveys)
Canadians (Political aspects)
Canadians (Political activity)
Canadians (Training)
Decision-making (Political aspects)
Amendments (Parliamentary practice) (Political aspects)
Education grants (Usage)
Education grants (Political aspects)
Author: MacIntosh, Constance
Pub Date: 09/22/2009
Publication: Name: Refuge Publisher: Centre for Refugee Studies Audience: General Format: Magazine/Journal Subject: Sociology and social work Copyright: COPYRIGHT 2009 Centre for Refugee Studies ISSN: 0229-5113
Issue: Date: Fall, 2009 Source Volume: 26 Source Issue: 2
Topic: Event Code: 980 Legal issues & crime; 280 Personnel administration; 290 Public affairs
Product: Product Code: E198620 Tenants; E198440 Refugees; 9103700 Immigration & Emigration; 9101226 Domestic Violence (Families); 9105113 Institutional Grants; 9103558 Exiles & Refugees NAICS Code: 92812 International Affairs; 92219 Other Justice, Public Order, and Safety Activities; 92311 Administration of Education Programs
Organization: Organization: United Nations. Office of United Nations High Commissioner for Refugees; United Nations. Office of United Nations High Commissioner for Refugees; United Nations. Office of United Nations High Commissioner for Refugees; United Nations. Office of United Nations High Commissioner for Refugees
Accession Number: 258439538
Full Text: Abstract

This report is an effort to address information gaps regarding how gendered claims are addressed by adjudicators at Canada's Refugee Protection Division of the Immigration and Refugee Board of Canada (the RPD). It looks at one specific type of gendered claim: persecution through domestic or intimate violence. The study considers all the RPD decisions from 2004 to 2009 and judicial reviews from 2005 to 2009 that were reported in the Quicklaw LexisNexis service. These decisions are analyzed both quantitatively and qualitatively.

This report finds adjudicators consistently identify domestic violence as a form of gendered persecution that can form a nexus to a convention ground. However, despite contrary directions from the Gender Guidelines, adjudicators often fail to recognize the social, cultural, economic, and psychological dynamics of domestic abuse as legally relevant for their assessment of state protection. There is a striking failure on this account when it comes to determining if it was reasonable to expect the claimant to seek state protection.

This report presents data on factors such as the rates at which adjudicators consider the adequacy of women's shelters and the responsiveness of local police to complaints. As well as identifying the frequency and grounds for which judicial reviews are granted, this report also presents a series of recommendations for reform. These recommendations identify where studies are needed, how the Gender Guidelines need reform to make them a helpful instrument, and how training and support for PRD adjudicators needs to be enhanced.

Abstract

Le present article tente de combler des lacunes documentaires tenant a la facon dont les juges de la Section de la protection des refugies (SPR) a la Commission de l'immigration et du statut de refugie du Canada traitent les revendications genrees. Un type particulier de revendication genree est examine : la persecution par la violence domestique ou conjugale. L'etude considere l'ensemble des decisions de la SPR de 2004 a 2009 et des commentaires judiciaires de 2005 a 2009 signales dans le service LexisNexis Quicklaw. Ces decisions sont analysees a la fois quantitativement et qualitativement. Il est constate que les juges identifient systematiquement la violence domestique comme forme de persecution genree, y trouvant motif a s'appuyer sur la Convention. Cependant, malgre les instructions contraires que renferment les directives sur le genre, les juges omettent souvent de reconnaitre la dynamique sociale, culturelle, economique et psychologique de la violence conjugale comme juridiquement pertinente pour leur evaluation de la protection de l'Etat. Il y a un echec frappant a cet egard quand il s'agit de determiner s'il eut ete raisonnable de s'attendre que le demandeur d'asile cherche la protection de l'Etat. L'auteure presente des donnees sur des facteurs tels que la frequence a laquelle les juges considerent le caractere adequat des refuges pour femmes et l'aptitude de la police locale a reagir aux plaintes. Tout en identifiant la frequence et les motifs pour lesquels les controles judiciaires sont accordes, l'auteure presente une serie de recommandations en vue d'une reforme. Ces recommandations indiquent ou des etudes sont necessaires, comment reformer les directives sur le genre pour en faire un instrument utile et comment la formation et le soutien des juges de la SPR doivent etre renforces.

Introduction and Overview of Findings

This report presents disturbing findings about how the refugee claims of women fleeing domestic violence are treated by adjudicators at the Immigration and Refugee Board of Canada. Although the Minister of Citizenship and Immigration is required to report to Parliament on the gendered impact of the Immigration and Refugee Protection Act, very little--if any--information has been synthesized and made public on this issue. The study underlying this report was designed to begin to fill this information gap--and the data presented in this report indicates that reform is urgently needed.

From 2004 to 2009 there were a total of 135 decisions by adjudicators of the Refugee Protection Division of the Immigration and Refugee Board CRPD") that were reported in the LexisNexis Quicklaw service where a woman sought protection from spousal violence. Three of those women were granted refuge. One hundred thirty-two claimants were rejected. The majority of these women were rejected on the basis that they had failed to rebut the presumption that their home state could protect them.

This report takes a quantitative and qualitative look at these decisions, with a particular emphasis on decisions that turned on state protection. It shows that RPD adjudicators are inconsistent in their adherence to the Gender Guidelines that were drafted by the Chairperson of the Immigration and Refugee Board (IRB) to guide the decision-making process when gender-based claims are made. There is some tendency to de-gender domestic violence claims by treating them like other cases of private criminality. In particular, adjudicators often fail to meaningfully engage with the complex social, cultural, and economic dynamics that impact on how victims of domestic violence are able to seek state protection. This puts the defensibility of these decisions into doubt. It raises questions about whether women are being returned to violence.

These findings are supported by the divergence between the factors the Gender Guidelines indicate RPD adjudicators are to take into account, and the factors which RPD adjudicators seem to treat as relevant in their decisions. These findings are also supported by the judicial reviews reported in this area. For the period of 2005 to 2009, a stunning 44 per cent of judicial review applications of decisions reported in LexisNexis Quicklaw where a woman was fleeing domestic violence from a spouse or ex-spouse were allowed by the Federal Court. (2) The majority of these decisions turned on finding that the RPD adjudicator's conclusion--that the woman's state could protect her--was not reasonable.

A successful judicial review application does not mean that a claim will ultimately be accepted on its merits: it only results in the claim being sent back to the RPD for a rehearing. However, it is evident that decisions are not being made properly, and so current practices are likely placing women at serious risk of being returned to violence and perhaps death at the hands of their prior spouse's.

This report makes many recommendations to try to bring the decision-making practices of RPD adjudicators into line with Canada's mandate to protect. It specifically calls upon the IRB to aggressively revisit their training practices with RPD adjudicators for hearing gendered claims, for Citizenship and Immigration Canada (CIC) itself to finally engage in a substantive review of the gendered impacts of protection decisions, and for the Gender Guidelines to be revised to provide more direction to RPD adjudicators in several problem areas.

Context and Scope of Study

Background: The Debate on How Best to Address Gendered Persecution (a Journey from Rapt Attention to Quiet Neglect)

Canada is a signatory to many international human rights instruments, including the 1951 Refugee Convention (3) and the 1967 Refugee Protocol. (4) The Refugee Convention contains a definition of "refugees." It defines them as persons who are outside of their state of nationality (or residency) and who have a well-founded fear of being persecuted on the basis of five enumerated grounds. These five grounds are race, religion, nationality, membership in a particular social group, and political opinion. The feared persecution must leave the person unable or unwilling to seek protection from their state of nationality. This definition is coupled with ah obligation--that signatory states are not to return "refugees" to the state where they face such persecution. Refugee protection is thus a surrogate level of protection, extended when a citizen is not protected by their home state.

This definition of "refugee," where protection is only offered if the persecution is linked to one of the five enumerated grounds, has long been acknowledged as incomplete. The incompleteness arises because it does not fully represent Canadian values and understandings of fundamental human rights. Canadians and Canadian law reject the legitimacy of targeting or leaving people vulnerable to harm based upon more than these five grounds; we also reject such targeting on the basis of grounds such as gender, age, and sexual orientation.

The potential gap between the definition of refugee and Canadian understandings of the fundamental human rights of women was identified as a problem by refugee adjudicators shortly after Canada's Immigration and Refugee Board (IRB) was established in 1989. Canada was not alone in identifying this problem. Since at least 1985, the United Nations High Commissioner for Refugees (UNHCR) has been urging states to recognize that women may be persecuted on the basis of gender, and that such women deserve the protection of the international community. (5) To further this objective, UNHCR released guidelines in 1991 that illustrated how the definition of "refugee" in international law could embrace gender-based persecution.

Canada struck committees in the early 1990s which were to study the issue and recommend a course of action. (6) Although major legislative reform of Canadas immigration and refugee regime was pending, and a fresh set of laws was to be passed in 1994, Canada chose not to revise its definition of refugees to explicitly recognize "gender" as a ground of persecution for which protection may be offered. Instead it was decided that the Chairperson of the IRB would exercise her statutory authority (7) and issue guidelines to direct adjudicators hearing refugee claims where the persecution is gender-related. This decision was inconsistent with the positions that Canada has taken in other contexts, where we have voluntarily and proudly explicitly bound ourselves by law to protecting women against wrongful treatment on the basis of their gender. For example, this commitment is reflected within the terms of our constitution and in both federal and provincial laws. (8) But in the context of refugee claimants, Canada, like the UNHCR, relies upon guidelines that indicate how existing laws can be interpreted to extend to women who suffer gendered persecution. (9)

These Guidelines were first issued in 1993, and then revised in 1996. (10) They were carefully drafted to sensitize adjudicators to how women may experience persecution in ways that differ from men, and to illustrate how such persecution may fall under the refugee definition. The Guidelines state:

The Guidelines address such issues as how to determine whether there is the necessary link between the gendered persecution and an enumerated ground, evidentiary issues that may arise in gendered claims, and other unique difficulties that may come up in the context of hearing a gendered claim. The Guidelines also describe three broad categories of situations where gendered persecution may arise. These include situations where women are persecuted due to their family relations (e.g., where the political opinions of other family members may be imputed to them, or they may be targeted as a route for getting at their family members), where women experience violence or severe discrimination due to their gender and their state does not or will not protect them, and where women are targeted due to failure to conform to "gender-discriminating religious or customary laws or practices." (11)

The then Chairperson of the IRB explained the decision to pursue guidelines, and not enact law, as reflecting the fact that "changing the definition unilaterally is a serious public policy issue" which "Canadians had to deal with in more depth." (12) This explanation suggests that the Guidelines were a potentially interim measure, pending more expansive consultations and analysis, and did not represent the end of the discussion on how Canada could best address claims involving allegations of gendered-based persecution. However, as discussed below, it does not appear that CIC has commissioned any substantive studies to help inform such conversations.

In late 1993 the Supreme Court of Canada confirmed that the enumerated ground of "membership in a particular social group" included groups defined by "gender." (13) One would have thought that this decision from the Supreme Court of Canada, coupled with the Guidelines, would together result in "gender" effectively operating as a sixth ground of persecution. However, many subsequent commentators have not found this to be the case. Instead, scholars have found that decision-makers sometimes struggle to make a plausible link between what they identify as gender-based persecution and the five enumerated grounds. (14)

The year 1994 was not the last time that Canada revisited its refugee legislation. There have been other instances when it would have been appropriate--and responsible--for Canada to have considered how the Guidelines were operationalized, and to rewrite either the Guidelines and/ or the legislation. Another major overhaul was undertaken in the late 1990s, leading up to the current statute, the 2001 Immigration and Refugee Protection Act (15) (IRPA). The National Association of Women and the Law (NAWL) submitted a brief to the Standing Committee on Citizenship and Immigration in which they made a series of recommendations about IRPA when it was in Bill form. (16) One of their recommendations read in part as follows: "We recommend that the refugee definition be amended to formally include gender in its own right, or as part of a larger open ended list of social groups." (17) This recommendation was not adopted. There is no suggestion that the decision to reject this recommendation was based on empirical research, an analysis of the Guidelines, or any other study into the operative outcomes and decisional processes of Canada's current approach to gendered claims. In these circumstances, the rejection of the proposal appears to have been arbitrary, or driven by factors other than an assessment of whether or how the Guidelines work, of whether or how we are returning women to persecution. Canada's neglectful approach to this area is further evidenced by the IRB having announced in 2002 that it would review the Guidelines in light of the changes brought about under IRPA, to assess whether it required modifications. (18) This review, however, appears to have never been performed, as there is no sign of it on CIC or the IRB's website. So we refused to change the law, while apparently completely failing to assess whether or not the law worked.

Citizenship and Immigration Canada does not appear to have ever followed up with an assessment of its decision to enact Guidelines, or to have otherwise analyzed the role the Guidelines play in decisions. Nor have the Guidelines been revised since 1996. This continuing gap in research and analysis exists despite the fact that there is a statutory obligation on the Minister of Citizenship and Immigration to report on such matters. In particular, the Minister is required by law to provide Parliament with a "gender based analysis" of the impact of Canada's immigration and refugee legislation on an annual basis. (19)

Although CIC does generate its annual reports, the reports are very general. They present little in terms of data or analysis on refugee claimaints. (20) For example, the only data regarding refugees included in the 2009 Annual Report to Parliament (21) are bar charts comparing permanent residents by immigrant class and sex from 2004 to 2008. From these charts we learn that of all men accepted as permanent residents from 2004 to 2008, 13 per cent became permanent residents due to obtaining some sort of "protected persons" status, and that of all women who became permanent residents, 11.5 per cent were landed within a "protected persons" category. The presentation of this minimal data does not constitute "a gender-based analysis of the impact of this Act" on women refugee claimants.

The 2009 Annual Report does refer to several studies of Canadas refugee resettlement programs which include gender-based considerations. These are obviously important research projects which are essential for determining how to best support the long-term success of woman refugees who are settled in Canada. However, these studies are fundamentally about how Canada develops its internal policies, not about the gendered impact of the Act itself. It is this later issue which must, according to statute, be analyzed. Based on the Annual Reports, it is unclear whether CIC is undertaking any substantive empirically-based analytic work on the gendered-impacts of IRPA. Calls to address this gap--of data and analysis of the data--have been made before. The most comprehensive and specific arguments are to be found in the 2006 Status of Women report entitled Gendering Canada's Refugee Process. (22)

It defies reason that CIC has interpreted its obligation to provide a "gender based analysis" of the impact of the Act as not including an analysis of gender-based claims. It is baffling that CIC has not reported on whether omitting "gender" from the legislation as a ground of persecution has had consequences for protecting women from persecution. This failure to report undermines Canada's ability to know whether we have made the right decisions, or whether we need to make fundamental changes. It undermines Canada's ability to know whether or how the RPD uses the Guidelines, and whether, in fact, the Guidelines adequately address and provide directions on the issues that arise in claims of gendered persecution. Although this report seeks to provide some information on these issues, it is no substitute for CIC fulfilling its obligations and certainly does not address the breadth of issues which require investigation.

Recommendation: That CIC fulfill its statutory obligation under IRPA s.94(2)(f) to include "a gender-based analysis of the impact of the Act" in its annual report to Parliament.

Recommendation: That CIC specifically perform a "gender-based analysis of the impact of the Act" on gender-based claims.

Whatever momentum was present in 1993, that led Canada to seriously consider how to best align Canadian values and approaches to fundamental human rights with the fact that women were seeking shelter from gendered persecution, appears to have been lost. The operating policy assumption must be that the Guidelines are effective, that the issue of how to adjudicate claims of gendered persecution no longer merits any concern. This report finds that this assumption is in error. It concludes that the Guidelines are often not being followed, and that many decisions are being made without proper regard to the social, cultural, economic, and policing context in which claimants may live. The failure to engage with the relevance of these contextual factors effectively de-genders the decision-making process. It has particularly strong consequences for assessments of whether the woman claimant acted reasonably in either not seeking police protection or not continuing to seek police protection in the face of having had her complaint disregarded by police officials.

Although policy and lawmakers have seemingly abandoned these issues, scholars have not. The merits of the Canadian government's decision to use non-binding guidelines, and to not enact laws, and the general question of how gendered claims are adjudicated, have been and continue to be the topic of considerable discussion. (23) Core critiques of Canada's current approach are that it trivializes gendered persecution, and makes the pervasiveness of gendered persecution invisible, and that anything less than legislation may be ineffective. (24) The core argument against making a legislative amendment, and instead continuing to work with just Guidelines, is that the legislation can already be, and is, read to include gender-based persecution--so there is simply no need to change anything. (25) A secondary argument is more concerned with the enumerated ground approach being inherently problematic because a list will always exclude many vulnerable groups. So the solution is either to abandon enumerated grounds altogether, or to argue for "women" being "a particular social group," thus creating precedents for interpreting this enumerated ground as encompassing persecution based on analogous grounds such as "age, sexual orientation, disability, etc." (26) These debates are, however, largely going on within a data vacuum, and would clearly benefit from information on how gender-based claims are being actually adjudicated. And, as noted above, there is very little information on this issue.

Recommendation: That CIC ensure all studies or reports which it produces that collect data or analyze gendered impacts be easily accessible by members of the public.

Data and Sample Set for Study

This study seeks to identify trends in decision making based primarily upon an empirical review. It is intended to provide baseline information on trends, as well as an initial analysis of those trends. A goal was to make it possible to begin to have an informed debate about how claims of gendered persecution are decided.

There are obviously many areas which require analyses of their gendered impacts, such as the experiences of sexual minorities within the claims process. However, this report does not set out to provide a comprehensive review of gendered claims. It only considers how refugee claims brought by women who flee alleged intimate or domestic violence from their spouse or ex-spouse have been adjudicated. This topic of analysis was selected because intimate violence has been identified as a "paradigmatic example" of gender specific abuse (27) which is also "the most prevalent yet relatively hidden and ignored forms of violence against women." (28)

The core data for this study is decisions by the Refugee Protection Division of the Immigration and Refugee Board. RPD adjudicators are the front-line decision makers who make the initial--and often only--assessment of a refugee claim. It is therefore essential to understand how the RPD adjudicates gender-based claims.

This study is limited to those decisions that were reported in the LexisNexis Quicklaw database. LexisNexis Quicklaw does not report all decisions of the RPD, but it does include a pool of decisions that is considerably larger than those which CIC posts in RefLex, their public and searchable online database. LexisNexis Quicklaw was searched for all decisions of the RPD that were reported as delivered in the years 2004 to 2009 (as of April 2010). The search terms were "domestic violence" and "intimate violence." The results of this initial search were then vetted on a case-by-case basis. Only decisions where the alleged persecution involved acts of violence committed against women by persons with whom they have or had an intimate spousal relationship were kept (i.e., all decisions where a person targeted a woman because she refused to start a relationship with that person, or where the violence was at the hands of another family member such as a father, were omitted). Although domestic violence may arise in many types of intimate or familial relationships, and men are also victims of domestic violence, these limits were put upon the decisions used for this report to try to minimize variables.) Ultimately, the total number of RPD decisions which were kept for the study was 135. As it turned out, all of the decisions involved heterosexual spousal relationships.

This study also drew upon some data which was produced pursuant to a Pro Bono Students Canada project for the Women's Legal Education Action Fund (LEAF) on judicial review assessments of RPD decisions involving domestic violence claims. (29) These decisions were identified by the students in a similar fashion: the LexisNexis Quicklaw database for Federal Court Trial Division decisions was searched for decisions which included the three terms "refugee," "gender," and "domestic violence." The pro bono students then reviewed the cases, eliminated those which were not domestic violence claims brought by women, and created brief summaries of key aspects of those decisions. This report draws upon some of their summary materials, for cases reported between 2005 and 2009 as of April 2010. The total number of cases is 89. The report author also reviewed all of the decisions where the judicial review was noted as allowed, to ensure that only reviews of RPD decisions involving spousal violence against women were included, and to extract relevant data. A list of all RPD and judicial review decisions that were reviewed can be found with the full version of this report that is posted at http:// www.nawl.ca.

Research Findings and Observations

The Big Picture: 98 Per Cent of Domestic Violence Cases Are Rejected by the RPD

As noted above, a total of 135 RPD decisions were reviewed. The expectation of the report author was that the decisions would demonstrate considerable variation in outcomes, reflecting facts such as the individual circumstances of woman claimants and the varying levels of social, cultural, policing or legal tolerance for domestic violence in different states. Instead they were essentially uniform.

Of the 135 decisions that were reviewed, 132 were rejected. This is a rejection rate of 98 per cent. However, in the years 2005, 2006, 2007, and 2008 the rejection rate was 100 per cent--no claims that were reported in LexisNexis Quicklaw were accepted by the RPD. The breakdown by year is provided in Chart 2.

Claims were denied most often because the woman claimant failed to convince the adjudicator that her home state could not protect her from her abuser (64 per cent). The second most common ground was that the claimant's story of having suffered from domestic violence was not found to be credible by the adjudicator (34 per cent). Other grounds included finding the claimant had an internal flight alternative (30) (11 per cent), that the claimant's delay in seeking protection was found to mean that they did not actually fear the alleged abuse (12 per cent), and the claimant's identity was not established (0.7 per cent). The grounds for declining claims--by the actual number of rimes the ground arose-are set out in Chart 3.

These decisions--that women were lying about their experience, or else that their state could in fact protect them--paint a picture that contradicts the trends identified by international bodies who study violence against women. For example, the United Nations General Assembly's 2006 assessment was that the "most common form of violence experienced by women globally is intimate partner violence" (31) and that "it is clear that violence against women remains a devastating reality in all parts of the world, and the implementation of international and regional standards to eradicate such violence is therefore ah urgent priority." (32) The UN essentially identifies domestic violence as a worldwide crisis that is being perpetuated by the failure of states to take effective action.

As a result of finding that their states could protect them from domestic violence, in the reviewed cases the RPD rejected individual woman claimants (and any accompanying children) over the last six years from: Bangladesh, Bolivia, Brazil, Bulgaria, Chile, China, Costa Rica, France, Germany, Ghana, Grenada, Guyana, Honduras, Hungary, Israel, Italy, Jamaica, Korea, Lithuania, Mexico, Namibia, Nigeria, Philippines, Portugal, Saint Lucia, Saint Vincent, Trinidad and Tobago, and the Ukraine. The United Nations has collected survey data from fourteen of these twenty-eight countries on domestic abuse. They found that the reported rates of women experiencing domestic violence are 40-47 per cent in Bangadesh, 27-35 per cent in Brazil, 25-26 per cent in Chile, 15 per cent in China, 9 per cent in France, 23 per cent in Germany, 10 per cent in Honduras, 32 per cent in Israel, 20.7 per cent in Korea, 42 per cent in Lithuania, 17-27 per cent in Mexico, 31 per cent in Namibia, 10-21 per cent in the Philippines, and 19 per cent in the Ukraine. (33)

The rates at which the RPD rejects claims on the basis that the woman's home state can protect her seem to make little sense in the face of this data on the general inability of these states to prevent the victimization of women by their spousal partners. It is hard to understand how a state could be considered to protect women if suffering spousal violence is a statistically common event. And, the Federal Court has found that RPD decisions were not reasonable, or otherwise could not stand, in 44 per cent of the decisions they reviewed from 2005 to 2009. (34) (This data is presented below.) This all raises questions about how decisions are being made.

This report now turns to trying to understand "the big picture." It presents more detailed information on trends and patterns in RPD adjudications of domestic violence claims, to identify what factors or evidence are taken as relevant and having weight. It also compares these trends to the directions in the Gender Guidelines. Although the report author had expected that the bulk of this study would consist of comparing positive and negative determinations, such a comparison would not be meaningful given that there were only three reported cases in the years under review where women were recognized as refugees. The report therefore focuses almost exclusively on the trends in negative determinations. Although some positive trends are identified, many are disturbing and suggest that Canada may be wrongfully returning women to danger.

Figuring Out What the Big Picture Means: Trends and Patterns in Adjudication

Trend: Domestic Violence Is Consistently Recognized as a Form of Persecution That May Found a Claim for Refugee Protection

One positive trend in the cases is that domestic violence is consistently recognized as a form of persecution that may found a refugee claim. This is consistent with the directions in the Guidelines. In their discussion of how an adjudicator ought to assess whether a feared harm constitutes persecution, the Guidelines direct that violence--"including sexual and domestic violence"--constitutes persecution when it is "a serious violation of a fundamental human right."

The survey of RPD decisions revealed no cases where the claim was denied on the basis that the alleged domestic violence did not constitute persecution. (See Chart 3.) When it came to the question of persecution, if the adjudicator believed the claimant's story, the adjudicator either described the violence which the woman had experienced, or simply asserted that the claimant had suffered from domestic violence, and then moved on to other steps of the analysis. This could suggest that the adjudicators have thoroughly understood and embraced the fact that domestic violence is always a violation of fundamental human rights, and therefore constitutes persecution. This outcome is an extremely positive one to document. It suggests that domestic violence--like forced prostitution, or state-sanctioned rape--has passed the conceptual threshold where its repugnance to fundamental human rights could still be debated.

The matter that appears to remain problematic for adjudicators, however, is what is required of states for them to be found to protect their citizens specifically from domestic violence. From the decisions of the RPD and judicial reviews from the Federal Court that were reviewed for this report (and which are discussed below), it is apparent that RPD adjudicators sometimes conflate the fact that protective legislation exists, or that police are undergoing gender-sensitivity training, with the conclusion that the legislation or training is in fact effective. Adjudicators' understandings must go from recognizing that domestic violence is repugnant to human rights, to recognizing that these rights continue to be violated for the purposes of human rights law until a state has taken meaningful steps to prevent its occurrence (and reoccurrence).

A secondary point is that all of the reviewed cases involved physical violence. The Guidelines do not expressly contemplate non-physical manifestations of domestic violence. This leaves a potential gap between the Guideline directions and contemporary human rights norms. As observed in a 2006 UN study of violence against women, in light of international human rights instruments, domestic violence is recognized as embracing a range of coercive acts which include sexual, psychological, and economic violence. The UN defines psychological violence as including "controlling or isolating the woman," and defines economic violence as including "denying a woman access to and control over basic resources." (35) The RPD's likely treatment of claims involving these forms of domestic violence is unknown. Given the emphasis in the Guidelines on physical violence, however, there is the potential for these other types of violence to not be recognized by an adjudicator as also being potential violations of fundamental human rights.

Recommendation: The IRB revise the Gender Guidelines to explicitly reflect contemporary human rights norms that define domestic violence as including sexual, physical, psychological, and economic violence.

Trend: RPD Assessments of State Protection Seldom Follow the Gender Guidelines' Directions

The Law on State Protection. As noted above, the reviewed intimate violence cases did not turn on whether the alleged domestic violence counted as persecution. They largely turned on a different aspect of the refugee definition. In particular, most turned on whether the claimant's home state is "unable" or "unwilling" to protect them. If their home state can protect them, then the claim is declined, as refugee protection is intended to be a form of surrogate protection. According to the Supreme Court of Canada, if a state has a functioning government, then it is presumed to be able to protect its citizens from private persecutors. As a result it falls on the claimant to rebut this presumption by presenting "clear and convincing" evidence of their state's unwillingness or inability to protect. (36) Where the persecution is at the hands of a private party--such as a spouse--the claimant is usually required to prove that they made efforts to seek protection from their home state, but that the protection "did not materialize." (37) This obligation on a claimant to have sought state protection only arises "in situations in which state protection 'might reasonably have been forthcoming" (38) and not when all that could be obtained is "ineffective protection." (39)

As well as receiving direct testimony from the claimant as to their experiences, adjudicators are provided with "country reports" to assist them in their assessment of state protection. These documents are compiled by a variety of governmental and non-governmental sources. For example, country reports are produced by the IRB and the United States Department of State, as well as by Amnesty International and Human Rights Watch. Country reports are drawn upon as information sources for matters that may be relevant for assessing refugee claims, such as whether and how allegations of police misconduct are addressed, whether there are laws in place that criminalize spousal violence, and conviction rates.

The Directions from the Guidelines on State Protection Do Not Seem to Play a Significant Role in Adjudicator Assessments. The Guidelines provide considerable elaboration on how to assess state protection in the context of gendered persecution. They flag a series of four factors that are relevant for determining whether the exception arises and it was reasonable for the claimant to not seek state protection. The Guidelines state:

The Guidelines similarly flag these contextual factors, as well as laws, as relevant for determining whether or not a claimant's home state can protect her. In particular, they require that in assessing state protection, "[t]he social, cultural, traditional and religious norms and the laws affecting women" ought to be assessed "by reference to human rights instruments which provide a framework of international standards for recognizing the protection needs of women." In other words, human rights instruments are to be used to identify the standards for assessing state protection of women. The Guidelines cite eight international instruments for adjudicators to draw upon. These are the Universal Declaration of Human Rights; International Covenant on Civil and Political Rights; International Covenant on Economic, Social and Cultural Rights; Convention on the Elimination of all Forros of Discrimination Against Women; Convention on the Political Rights of Women; Convention on the Nationality of Married Women; Convention Against Torture and other Cruel, Inhumane or Degrading Treatment or Punishment; and Declaration on the Elimination of Violence Against Women.

Based on the cases reviewed for this report, these directions have not been followed with any consistency by the Refugee Protection Board in its negative decisions that turned on state protection. Instead:

* protection standards from human rights instruments do not seem to inform decision-making;

* RPD adjudicators often seem to be selective in their assessment of protective legislation and policing support--while weight is always given to evidence that the state can protect, weight seldom seems to be given to contrary evidence;

* RPD adjudicators do not appear to treat social or cultural factors as relevant for determining whether it was reasonable for the woman to seek state protection; and

* Economic and religious factors do not appear to be taken meaningfully into account in determining if it was reasonable for the woman to not seek police intervention.

The evidence for these findings, and related ones, is discussed below.

Protection Standards from Human Rights Instruments Do Not Seem to Inform Decision Making. In the negative decisions that were reviewed, international human rights instruments were not used to identify standards or thresholds for protecting women. This constitutes a failure to follow the Gender Guidelines. The pragmatic result of this is the likely failure of adjudicators to appreciate the distinct character of women's protection needs when their persecution is based in gendered discrimination.

Of the eighty-six claims where the presumption of state protection was not rebutted, only four sets of reasons were identified as referring to international human rights instruments. (40)

In two of these cases the adjudicator observed that relevant human rights instruments had been ratified. The adjudicator did not, however, consider whether the states had implemented their obligations. The implication seems to be that the fact of signing such instruments means that the state should be deemed safe. Ina third decision the adjudicator concluded that, as the woman's home state had signed the treaty on the Elimination of Discrimination against Women, one could presume domestic compliance and implementation.

These are dangerous leaps of logic. They may be reasonable assumptions to make in the case of treaty obligations generally, but the data shows that such an assumption is not reasonable in the case of treaty obligations to protect women. The United Nations and treaty monitoring bodies have documented the all-too-common gap between states committing to protect women from violence, and actually making good on those commitments.

In only one case did the adjudicator consider reports by treaty bodies on compliance. Here the adjudicator found that commentators were concerned that many treaty obligations had not been met. However, the adjudicator nonetheless found that the presumption of state protection was not rebutted.

Recommendation: The IRB ensure that RPD adjudicators are provided with reports from human rights treaty bodies on state compliance with their treaty obligations to protect women.

RPD Adjudicators Tend to Only Partially Assess the Impact of Domestic Laws on Women Seeking Protection from Domestic Violence. The Gender Guidelines clearly call on adjudicators to consider whether there are laws that protect women from domestic violence. In most decisions, adjudicators did take note of whether protective legislation existed (86 per cent). However, in most cases this was the end of the analysis. Adjudicators only commented upon the effectiveness of such laws in 29 of the 86 cases. (See Chart 5.) This is a serious problem. It may usually be reasonable to assume that, if a state enacts protective laws, such laws will be enforced. However, this assumption is not a reasonable one to make in the case of domestic violence laws. In many states women have suffered and continue to suffer from deeply ingrained patterns of discrimination as well as cultural or social norms which tolerate violence against them, and which undermine the effectiveness of legislation. Once again, the United Nations has documented the persistent gap between enactment and the effective implementation of laws that protect women from domestic violence in many states. (42) Further inquiry is always required.

The Gender Guidelines are sensitive to this issue. As noted above they refer to human rights instruments to identify the standards for women's protection needs. These instruments are clear that women are only protected if laws are effective.

One of these instruments is the Convention on the Elimination of All Forms of Discrimination Against Women. Article 3(c) of the convention requires the establishment of "legal protection of the rights of women" and "the effective protection of women against any act of discrimination."

Another instrument which decision makers are directed to for the standard for women's protection needs is the Declaration on the Elimination of Violence Against Women. Article 4(d) requires states to develop domestic legislation which provides "just and effective remedies for the harms" suffered by women who are subjected to violence.

These instruments indicate that state protection is only present where there are effective state responses. The proper threshold, then, if the Guidelines are followed and domestic violence claims are not to be de-gendered, is to inquire into the effectiveness of state measures, not merely to assess whether measures exist in law. This requires assessing state legislation on an operational scale. If laws are not enforced, or are ineffective or inaccessible due to requirements such as there being an independent witness to the violence, then women are not in fact protected by their states.

Recommendation: The IRB revise the Gender Guidelines to describe the thresholds for protecting women that are identified in the international human rights instruments that are referred to in the Guidelines. Ata minimum, that the Guidelines should incorporate the standard from the Convention on the Elimination of Ali Forms of Discrimination Against Women that the state provide "effective protection" and "just and effective remedies."

Recommendation: The IRB ensure RPD adjudicators are trained to interpret and apply the standards from the international human rights instruments that are referred to in the Gender Guidelines to assessments of state protection of women.

Recommendation: The IRB enhance their training programs for RPD adjudicators to ensure that adjudicators are sensitive to the distinctions in assessments of state protection between claims where the persecutor is a spouse or ex-spouse, and claims where the persecutor is involved in other sorts of criminal activity.

Although RPD adjudicators are taking note of whether there is protective legislation in the majority of the cases, this finding cannot stand in for an analysis of whether the state is in practice willing and able to protect women from domestic violence. "The potential of laws on violence against women remain unfulfilled if they are not effectively applied and enforced." (43)

A related trend is that evidence that protective laws exist is always given weight in assessing whether the presumption of state protection was rebutted. However, in the three cases where no laws that criminalized or made domestic violence actionable existed, the presumption of state protection was still not rebutted. Such evidence does not seem to attract any weight. This raises the spectre that some adjudicators are being selective in their use of country report evidence, instead of fair-handed.

RPD Adjudicators Are Selective in Giving Weight to Police Responsiveness to Domestic Violence Victims. There is a similar pattern of inconsistency in terms of adjudicators considering documented police attitudes towards women who seek protection. Although some adjudicators observed whether police were usually responsive, or reluctant to assist domestic violence victims, in 60 per cent of the decisions there was no consideration whatsoever of documented police attitudes towards women who sought protection from domestic violence. It is impossible to assess whether a claimant acted reasonably in not seeking police protection if the adjudicator does not consider whether the police are likely to refuse to assist the claimant. However, in 60 per cent of the cases where the presumption of state protection was not rebutted, this unwarranted jump in logic was made.

In the claims where documentary evidence on police attitudes to alleged victims of domestic violence was considered, adjudicators found that police were usually responsive in 23 per cent of cases and reluctant to intervene in 31.5 per cent of cases; they also found there were radically different documented conclusions on police attitudes in 14 per cent of the cases. In several of the cases (34 per cent), the adjudicator found that police attitudes were improving, but provided no assessment of whether the police were in fact actually responsive: the fact of improvement seemed to stand in for a conclusion on this point. This could suggest that adjudicators were basing their assessment on the "good will" of states, not whether women were protected. In all five of the cases where there was radically contradictory evidence on police attitudes, the adjudicators preferred the evidence that police were responsive, effectively finding the claimant's own undisputed evidence in three of these five cases that their own attempts to seek police assistance had been rebuffed to have no weight.

In practice, where police were documented as being responsive, this was taken as evidence that the presumption of state protection was not rebutted. However, when police were documented as not being responsive, or as "improving," or where there was some evidence of a lack of responsiveness, this documentary evidence does not appear to have carried any weight in the decision that the state could protect. This is despite the fact that police attitudes are one of the factors that the UN identifies as causing otherwise adequate protection regimes to offer ineffective protection in practice.

The trend in some RPD decisions to selectively draw upon positive evidence of state protection--and to ignore contrary evidence--is also commented on in judicial review decisions. These are discussed below.

RPD Adjudicators Do Not Appear to Treat Social or Cultural Factors as Relevant for Determining If It Was Reasonable for the Claimant to Not Seek State Protection. The negative decisions were also reviewed to see what part social or cultural factors played in finding the presumption of state protection was not rebutted. In particular, they were flagged if and when the adjudicator cited documentary evidence in their decision or when he/she made statements about the following social or cultural contextual issues:

* if there were cultural or societal attitudes of violence against women

* any observations on how cultural or social norms relate to risks of domestic violence

* whether domestic violence was a "serious" or "widespread" problem

* whether domestic violence is considered a "private matter" or "family affair" or source of shame

* whether women were reluctant to report (or did not report) domestic violence

* whether the claimant alleged police apathy or reluctance to act, as well as the claimant's own experiences in seeking police assistance.

Of the 86 decisions that were reported from 2004 to 2009 where a ground for declining the claim was that the presumption of state protection was not rebutted, the adjudicator only considered if there were cultural or societal attitudes of violence towards women, or whether domestic violence was considered a "private matter" or "family affair" in 17 cases, a mere 20 per cent of decisions. (See Chart 7.) Most of these cases were released in 2004 (8 cases) and 2005 (5 cases).

Although in the majority of cases there was some sort of comment about domestic violence being prevalent in the woman's home state, this observation never seemed to carry any weight for assessing state protection. One would think that if domestic violence was found to be a "serious" and "widespread problem" that this evidence would go some way towards rebutting the presumption of state protection. However, these observations do not appear to be taken as relevant in the final analysis.

Where adjudicators drew in other social or cultural factors--which they did at a very low rate--these too did not seem to affect the assessment of whether it was reasonable for the claimant to have assumed that their state would not protect them and so did not seek state protection, or failed to continue to pursue police protection in the face of having their requests for assistance rebuffed by police officers.

Given that most adjudicators made at least a minimal reference to societal conditions, it is clear that most understand the need to draw upon these factors. They do not, however, appear to know how to use the information--or to know what information is relevant--in making their decision. The problem is perhaps one of training and support. RPD adjudicators cannot be assumed to be sociologists or political scientists--the jump between identifying socio-cultural data and actually determining its legal relevance is not an easy leap to make. However, if the Guidelines are to be respected, and taken as good guidance for determining gendered claims, then this gap must be closed.

Recommendation: RPD adjudicators be specifically trained and supported in understanding the relevance of socio-cultural factors for assessing the reasonableness of women not seeking--or failing to seek with persistence--police assistance.

Economic and Religious Factors Do Not Appear to Be Taken Meaningfully into Account. As well as directing adjudicators to consider how social and cultural factors are relevant for determining whether it is reasonable for a claimant to seek state protection, the Guidelines also direct adjudicators to consider economic and religious factors. Religious factors were not mentioned in any of the 86 decisions.

As to economic factors, a few were referred to briefly in fewer than half of the decisions. In particular, in 38 of the 86 decisions (44 per cent), adjudicators identified whether or not there were shelters in the claimant's home state for women fleeing domestic violence. However, identifying the existence of shelters tended to be the end of the analysis. There were only 13 decisions where the adjudicator referred to documentary material on adequacy. In all of these cases, the adjudicator found that either there were no shelters or that the shelters were very insufficient to meet demand. However, these findings--that women may have no where else to go-appeared to have played no role in the conclusion that the claimant ought to have sought (or sought more aggressively) police intervention.

The UN Committee for the Elimination of Discrimination against Women has specifically expressed concern over women's shelters in signatory states being inadequately financed and their operations unmonitored. (44) One cannot assume that the reported existence of shelters translates into shelters that are operating and effective.

Also of note is the tendency of adjudicators to discuss the existence of private or NGO shelters, or shelters operated by international aid agencies, within their discussion of whether there is evidence of state protection, even though it is obvious that such private or international initiatives most like arise due to a lack of state shelters. This conflation of state initiatives with non-state initiatives is misleading and distorts the analysis of state protection. The Gender Guidelines specifically address this issue, and so apparently are once again not being followed.

Recommendation: In assessing state protection, RPD adjudicators be cautioned to follow the Gender Guidelines and distinguish between protective measures that the state offers, and those which are provided by non-state entities.

There was little consideration of any other economic circumstances which may be relevant in assessing whether it is objectively unreasonable for the claimant to have sought state protection through police intervention. The only other example of note is that in a few decisions, the adjudicator referred to country documents that indicated women were economically dependent on men. This was described as part of the general context in 5 of the 86 cases, but appeared to have no impact on the analysis of state protection. In the remaining 81 cases no consideration whatsoever was given to how economic considerations may impact on the behaviour of abused women, and the analysis of whether it is objectively unreasonable for them to seek state protection.

In one case, (45) the claimant put forward economic evidence as to why she had only sought police protection intermittently, and had also continually returned to her abusive spouse over a twenty-four-year period. Her uncontested evidence was that she returned--and did not seek much police intervention--because she could not support her children by herself. Her evidence was also that her family refused to provide her with food and housing, and told her instead to return to her spouse. We see here how societal attitudes and economic needs impact on women's ability to take action. Although the adjudicator observed the woman behaved in accordance with housing needs, he nonetheless found the claimant "failed to make a diligent effort to obtain state protection," and so the presumption of state protection was not rebutted. The adjudicator did not consider how the claimant could survive without the spouse that she was dependent upon, or where she was to live.

As mentioned above, there were five cases included in the review where adjudicators made general links between economic factors and how abused women may behave. For example, adjudicators referred to documentary evidence that women were reluctant to seek the prosecution of their abusive spouses due to financial dependence upon those spouses. Adjudicators also observed that even when the state issued protection orders, country condition reports indicated that women would let their spouses back into the family home--and not report the violation of these orders--due to financial dependence upon their spouses. (46) However, these careful and sensitive observations about how economic dependency affect abused women's decisions about seeking state protection did not seem to play a role in assessing whether the women had acted "reasonably" in their decision to not seek state protection, or to not seek further assistance to exclude their abusive husbands when they re-entered the family home.

Recommendation: RPD adjudicators be trained and supported in understanding how a claimant's economic and religious context may be legally relevant for assessing the reasonableness of women not seeking--or failing to seek with persistence--police assistance. In particular, adjudicators should receive direction on how economic dependency may restrict women's ability to leave a violent relationship, and to be able to meet basic economic needs if their male spouse is removed from the home.

Snapshot: An Example from Grenada. In the unsuccessful 2006 claim of Lewis v. Canada (MCI) (47) there was a majority and a dissenting set of reasons. The decision turned on whether it was reasonable for the claimant to have failed to go to the police, and the majority and dissent part on the relevance of social and economic factors. Although both observed that shelters existed, only the dissenting adjudicator considered the actual availability of shelters. Similarly, only the dissenting adjudicator referred to evidence on how financial dependency plays into the ability of women in the claimant's position to pursue charges. The claimant also gave evidence of other women whose attempts to seek police protection had been rebuffed, and stated she believed her ex-partner would carry out his threat to kill her if she went to the police. Given the economic, social, and policing context, the dissenting adjudicator found it was not unreasonable for the claimant to not seek police protection. The majority panel, however, rejected her claim. They wrote:

The majority explicitly and blatantly failed to follow the directions in the Gender Guidelines. They correctly identified that they were being asked to consider how the social and cultural context affected the reasonableness of the claimant failing to seek state protection. However, they refused to engage in the required analysis. It appears that they did not understand how it could be legally relevant. The majority commented a second time on these broader contextual factors in their concluding comments on state protection. They wrote"

The majority found the first reason (police disinterest) had been gradually addressed. However, they made no comment about whether the second reason (economic survival) still persisted. So although the majority clearly linked women's failure to seek police protection to restrictions created by social and economic norms, the majority nonetheless failed to consider whether "economic survival" would make it reasonable for a woman to fail to seek police assistance. The failure to consider the relevant of gendered economic dependency once again resulted in the de-gendering of the analysis.

Recommendation: The IRB ensure RPD adjudicators are trained in how social, cultural, economic, and religious factors can be relevant for assessing whether the presumption of state protection has been rebutted.

Trend: RPD Adjudicators May Be Insensitive to How Women May Behave If They Are Victims of Spousal Violence

The Gender Guidelines explicitly flag the fact that "[w]omen who have been subjected to domestic violence may exhibit a pattern of symptoms referred to as Battered Woman Syndrome ..." The Guidelines refer to the Supreme Court of Canada decision, R v. Lavallee, (50) for understanding how abused women may behave. The Guidelines include the following quote from Lavallee:

[the myth is that e]ither she was not as badly beaten as she claims, or she would have left the man long ago. Or, if she was battered that severely, she must have stayed out of some masochistic enjoyment of it.

This quote highlights two "myths" about abused women that adjudicators must be careful to avoid. The first myth is to conclude that any woman who alleges that she stayed in an abusive relationship must be lying. The second myth is to conclude that if the woman was not lying, then the abuse was not actually experienced by her as abusive, and therefore does not deserve recognition as abuse, or to be otherwise recognized as legally relevant.

The Gender Guidelines go on to point out that one of the manifestations of Battered Woman Syndrome is that the abused women is "reluctant" to disclose the existence of the abuse. In the Lavallee case, the woman who was found to suffer from Battered Woman Syndrome demonstrated this pattern by failing to seek police assistance and by lying to doctors about the source of her injuries. The Supreme Court cautioned that in the case of abused women, such inaction and/or deceit should not necessarily impugn the credibility of their delayed claim of abuse, or their general credibility.

Several cases illustrate adjudicators buying into and dismissing claims based on these myths, instead of identifying and rejecting them. Some of the language from these cases is highlighted in the box below. It is clear from these extracts that adjudicators may see women as making choices to permit abuse to continue: either through remaining in a relationship or else by failing to go to police. In either case the implication is that their situation is their own fault--and therefore the woman is unworthy of asylum.

Snapshot: Wrongful Assumptions about the "Reasonable" Abused Woman.

* [she] "has made no serious efforts to break off the relationship" (51)

* "she might have been an abused wife but she never did anything to help herself" (52)

* "... together with the fact that she went back to him in 2002, is hardly a vigourous pursuit of protection" (53)

* "It is not sufficient to go to the police only on one occasion and then to abdicate her responsibility by not reporting the [subsequent] alleged threats" (54)

* "In terms of tools available to women who are victims of violence, the panel might add ... that there is a divorce law allowing spouses to end their relationship once and for all" (55)

* "it has to be noted that the claimant's erratic relationship with the perpetrator always ended in the claimant reconciling with him." (56)

* "Three rapes, six moves and the loss of three jobs due to the behaviour of her assailant, xxxx xxxxx, did not convince the claimant to lay a complaint or to seek help or assistance from the police or a women's right's organization in Costa Rica because, according to her, protection in that country takes a long time to obtain and is unreliable. She also feared that xxxxx would be imprisoned for two months and then get out of jail, kill her and then commit suicide: this became her excuse for not availing herself of the protection of the Costa Rican authorities" (57)

* "... she backed away from pursuing the charges because xxxxx said he was going to kill her ... Because she refused to press charges the claimant did not fully discharge her obligation to seek state protection" (58)

The matters of RPD adjudicator assessments of credibility and what sort of behaviour one can reasonably expect of women who have been abused by their spouses are taken up again later in this report. In particular, section 3(c) presents information on how often RPD decisions are found to be unreasonable due to the adjudicator's assessment of the claimant's credibility.

Recommendation: That the IRB work to better sensitize RPD adjudicators to the myths of domestic violence; In particular, to develop their understanding of what is "reasonable" or expectable behaviour for abused women, including remaining in or returning to abusive relationships and failing to seek police or medical assistance.

Judicial Reviews of RPD Decisions Confirm Disturbing Trends

The Judicial Review Rate for Negative PRD Decisions on Domestic Violence Claims Is 44 Per Cent RPD decisions are usually expected to stand. RPD adjudicators are presumed to possess specialized expertise. They are also seen as best placed to assess many aspects of the evidence and, of course, the credibility of the claimant. Nevertheless, an unsuccessful claimant may bring an application to the Federal Court, seeking to have a decision judicially reviewed. Any judicial review of factual findings--such as whether a claimant's state can protect her, or whether the claimant's story was credible--will be performed with considerable deference to the adjudicator's judgment. Such findings are reviewed against a standard of "reasonableness." "This standard is met when the "decision falls within the range of possible, acceptable outcomes which are defensible in respect of the facts and law." (59) The reviewing judge centrally asks whether the RPD adjudicator justified their decision, and whether the reasons are transparent and intelligible. So decisions stand where they are "defensible," and one can see how they were reached--not because a reviewing judge finds that they were correct, or because the reviewing judge would have made the same decision. On the other hand, where the claimant argues that the RPD adjudicator made a legal or process error--such as applying the wrong legal test or acting unfairly--the reviewing judge will ask if the adjudicator got these matters right.

A review of the LexisNexis Quicklaw Federal Court database from 2005 to 2009 reveals that there were 89 judicial reviews for negative RPD decisions where the claimant was fleeing domestic violence. (See Chart 11.) On average, for this five-year period, a shocking 44 per cent of these reviews were allowed. That is, in 39 of the 89 reported decisions, the Federal Court found that the RPD adjudicator either had the law or an element of procedural fairness wrong, or else had reached indefensible conclusions on factual matters.

Canada's self-proclaimed (and commendable) policy on refugee hearings is that the "system focuses on getting the decision right at the first level, with highly qualified and well-trained decision-makers ..." (60) The judicial review rate indicates that RPD adjudicators are not meeting Canada's policy objectives. Intervention is clearly required.

Federal Court Judges Condemn Some RPD Decisions for Only Giving Weight to Country Reports That Indicate a Woman's Home State Is Safe

The reasons for permitting judicial reviews in these cases are also troubling. In 77 per cent of the cases, the reviewing judge allowed the judicial review on a basis that the adjudicator's finding on state protection did not meet the reasonableness standard. (See Chart 12.) In the majority of cases, this was because the adjudicator failed to address evidence that was placed before him or her that the state could not protect the claimant. That is, in assessing whether the presumption of protection was rebutted, the adjudicator did not consider the presented evidence showing that the woman's state could not protect her.

Often the Federal Court judge's description of why the adjudicator's decision was not defensible is presented in a fairly neutral fashion. For example, in one case the judge simply wrote: "The RPD failed to explain its selective reliance upon the documentary evidence and specifically failed to deal with the evidence that directly contradicted its finding that protection was available to women in Mexico." (61)

However, the frustration of the Federal Court with the assessments of state protection by RPD adjudicators is in some instances quite palpable. For example, in a 2009 case where a reviewing judge found the RPD assessment that St. Vincent and Grenadines could protect a claimant from domestic violence was unreasonable, the judge wrote: (62)

5 Taken at its face value, the decision appears to be reasonable. This Court is supposed to show deference to the RPD panels who allegedly have greater expertise in country conditions than the Court itself. However there comes a time when it becomes obvious that deference should be earned, particularly when the Panel apparently pays no attention to the cases coming out of this Court which specifically deal with St. Vincent and the Grenadines. The analysis of country conditions was clearly a pro-forma one, or what Madam Justice Snider called in Alvandi v. Canada (Minister of Citizenship and Immigration), 2009 FC 790, a "cookie-cutter analysis."

6 This Court does not sit in a de novo appeal and so cannot do its own country analysis....

7 .... there are a great number of cases where judicial review has been granted on the basis that findings that there is state protection in St. Vincent and Grenadines were unreasonable. Without putting too tine a line on it, many of the women appear to have been in generally similar situations. See for instance: [list of eight cases where judicial review was allowed on state protection]

There are many other cases where the Federal Court similarly found the RPD adjudicator had selectively relied only upon the evidence that the state could protect, and failed to meaningfully engage with evidence that demonstrated the claimant's home state could not protect her. (63)

Reviewing Judges Find RPD Assessments of Credibility Ma), Rest on False Assumptions about Domestic Abuse

Perhaps even more surprising are the judicial reviews that were allowed on the basis of the adjudicator making unreasonable conclusions about the credibility of the woman's story. Credibility was the second most frequent reason in the studied decisions for the RPD to decline claims, at 34 per cent. An RPD adjudicator's assessment of claimant credibility may turn on the adjudicator's understanding about what is not reasonable behaviour for an abused woman. For example, in one case (64) a woman's claim that she was abused by her spouse was disbelieved by the RPD adjudicator in part because she had remained in her home with her alleged abuser for a year after receiving a visa to travel to Canada. The reviewing judge found the RPD's conclusion on credibility to be unreasonable, writing:

Similar comments are scattered through the judicial review decisions, of judges determining that credibility findings were unreasonable because they were based on claimants being unable to explain why they stayed in an allegedly abusive relationship, or why their spouse had become violent toward them. (65) The observations in these cases marry with the conclusions reached above in the discussion of RPD adjudicators being insensitive to how abused women may behave. This problem is clearly one of education, of supporting adjudicators to develop a better understanding of how behaviour which may otherwise seem incredible is in fact not unusual or unreasonable for a woman in an abusive spousal relationship.

Reviewing Judges Often Link RPD Errors to a Failure to Follow the Gender Guidelines

In one-third of the cases where the judicial review was allowed, the reviewing judge observed that the adjudicator had failed to follow and apply the Gender Guidelines. This failure, in turn, was typically linked to why the adjudicator's conclusion on credibility or state protection was unreasonable. This suggests that, if the Gender Guidelines are actually considered and properly applied, then decision making will improve and the judicial review rate may drop.

Recommendation: The IRB undertake an aggressive review of their training and support programs for RPD adjudicators hearing gender-based claims. There should be an emphasis on the areas which are regularly raising concerns with the Federal Court. These include receiving and assessing the credibility of testimony from persons who may have been victims of domestic abuse, and the necessity to assess all relevant evidence regarding state protection--not just evidence that the state can protect.

Recommendation: The IRB deliver a revised training and support system for RPD adjudicators for assessing gender-based claims.

Recommendation: The IRB track, analyze, and publish data on positive and negative trends identified through judicial reviews of its decisions on domestic-violence-based claims.

Recommendation: Trends identified through judicial reviews be used to identify where more direction is needed in the Gender Guidelines, and where adjudicators need more training and support.

Concluding Comments

The Guidelines receive inconsistent treatment by adjudicators. Although adjudicators properly assess domestic or intimate violence to be a form of persecution, the gendered analysis is often lost when it comes to assessing state protection. In particular, adjudicators tend to conflate the fact that states have enacted protective legislation with the finding that there is protection for the claimant. RPD adjudicators also seem to fail to meaningfully engage with how the social, cultural, economic, and policing context are relevant for evaluating how abused women may be able to engage with the state apparatus and policing services.

Given the rates at which domestic violence claims have been dismissed by the RPD (almost 100 per cent in the study set), the consistently high rates of successful judicial reviews over the past five years (averaging 44 per cent in the reviewed cases), and the scathing commentary on some RPD assessments of state protection and claimant credibility, as well as observations on the failure to follow the Gender Guidelines, it appears that many RPD adjudicators have not been adequately trained and supported to make reasonable decisions on claims involving domestic violence. These findings would seem to indicate that it is more likely than not that Canada is indeed returning women to danger.

This report does not assert any conclusions about whether Canada would better meet its protection obligations if it was to amend the refugee definition to explicitly include "gender" as a ground of persecution. However, it does show an express need for adjudicators to receive better training for hearing gendered claims, and that the Gender Guidelines themselves require considerable revision if they are to serve as an effective tool for assisting adjudicators. Pivotally, revisions are required to bring them in line with contemporary human rights instruments, and to provide better and more fulsome direction on a number of issues, such as how adjudicators are to use social and economic norms in their assessments of state protection. This indicates the pressing need for the Guidelines to not merely be revised on an issue-by-issue basis, but to be put through a general review process. Obviously, after such a review (and revisioning), there needs to be a process in place to ensure that the Gender Guidelines are regularly revisited. Given the findings of this report, they ought to be regularly reviewed for revisions both to reflect changing norms and law, as well as to reflect areas where adjudicators are found to need more direction.

Recommendation: The Gender Guidelines be put through a comprehensive review to revise them to reflect contemporary law and norms, and to provide better guidance in the areas where RPD decisions seem to falter.

Recommendation: CIC instantiate a process for reviewing and revising the Gender Guidelines on an on-going basis.

NOTES

(1.) Canadian Immigration and Refugee Board, "Women Refugee Claimants Fearing Gender-Related Persecution" (Guidelines issued by the Chairperson of the Board in accordance with subsection 65(3) of the Immigration Act) (Ottawa: Immigration and Refugee Board, 1993); "Women Refugee Claimants Fearing Gender-Related Persecution: Update" (Ottawa: Immigration and Refugee Board, 1996) [Gender Guidelines].

(2.) This timespan was selected to reflect the approximate period of time for when RPD decisions rendered in 2004 to 2009 might be judicially reviewed.

(3.) Convention relating to the Status of Refugees, 28 July 1951, 189 U.N.T.S. 150 (entered into force 22 April 1954 [Refugee Convention].

(4.) Ibid.

(5.) For example, Refugee Women and International Protection, EXCOM Conclusion No. 39 (XXXVI) (1985), paras. (b) and (k).

(6.) Nicole LaViolette, "Gender-Related Refugee Claims: Expanding the Scope of the Canadian Guidelines" (2007) 19:2 Int'l J. Refugee L. 169 at 174.

(7.) S.65(3) Immigration Act, R.S.C. 1985, c.I-2, continued after 2002 under s. 159(1) (h) Immigration and Refugee Protection Act S.C. 2001, c.27 [IRPA].

(8.) For example, Section 15 of the Canadian Charter of Rights and Freedoms prohibits discrimination on the basis of sex, as do the provincial and federal Human Rights Codes.

(9.) Gender Guidelines, supra note 1.

(10.) Ibid.

(11.) For a more detailed discussion, see Constance MacIntosh, "When 'Feminist Beliefs' Became Credible as 'Political Opinions': Returning to a KeyMoment in Canadian Refugee Law" (2005) 17:1 CJWL 135; N. Mawani, "Introduction to the Immigration and Refugee Board Guidelines on Gender-Related Persecution" (1993) 5:2 Int'l J. Refugee L. 240

(12.) Mawani, ibid. at 242.

(13.) Canada (AG) v. Ward, [1993] 2"SCR 689 at 739.

(14.) See for example, Chantal Tie, "Sex, Gender and Refugee Protection in Canada under BiU C-11: Are Additional Protections Required in Light of In re R-A?" (2001) 19:6 Refuge 54; Melanie Randall, "Refugee Law and State Accountability for Violence Against Women: A Comparative Analysis of Legal Claims Based on Gender Persecution" (2002) 25 Harv. Women's L.J. 281 at 290-292. For a discussion of the limited fashion in which "gender-based" persecution has been defined by Canadian decision makers, see Laviolette, supra note 6.

(15.) S.C., 2001, c.27.

(16.) NAWL, Brief on the Proposed Immigration and Refugee Protection Act (Bill C-11), Apri1 2001.

(17.) Ibid. at 28. The Recommendation reads in whole: "We recommend that the refugee definition be amended to formally include gender in its own right, or as part of a larger open ended list of social groups. In any case, 'social group' should be defined in Bill C-31 as explicitly including gays and lesbians."

(18.) See "Update" online: http://www.irb.gc.ca/Eng/notavi/Pages/cgi.aspx.

(19.) IRPA, supra note 7, s.94(2)(f).

(20.) See comments in Catherine Dauvergne, Leonora Angeles, & Agnes Huang, Gendering Canada's Refugee Process (Ottawa: Status of Women Canada, 2006) at 19-21.

(21.) Citizenship and Immigration Canada, "Gender-Based Analysis of the Impact of the Immigration and Refugee Act," Section 5 in Annual Report to Parliament on Immigration, 2009 (Ottawa: Citizenship and Immigration Canada, 2009).

(22.) Dauvergne, Angeles, & Huang, supra note 20 at 18-24.

(23.) See for example Audrey Macklin, "Refugee Women and the Imperative of Categories" (1995) 17 Hum. Rts. Q. 213; Randall, supra note 14; MacIntosh, supra note 11; and Dauvergne, Angeles, & Huang, supra note 20.

(24.) For a description and critical discussion of these arguments, see Macklin, "Refugee Women" ibid. at 256-261.

(25.) See for example, Deborah Anker, "Refugee Status and Violence Against Women in the 'Domestic' Sphere: The Non-State Actor Question" (2001) 15:39 Geo. Immigr. L.J. 391.

(26.) Macklin, "Refugee Women" supra note 23 at 262.

(27.) Audrey Macklin, "Cross-Border Shopping for Ideas: A Critical Review of United States, Canadian, and Australian Approaches to Gender-Related Asylum Claims" (1998) 13 Geo. Immigr. L.J. 25 at 56.

(28.) UNICEF, "Domestic Violence Against Women and Girls" (2000) 6 Innocenti Digest at 2.

(29.) This work was carried out by Aaron Martens and Anita Kim, with their report submitted to LEAF on April 3, 2010.

(30.) The term "internal flight alternative" refers to situations when there is somewhere in the claimant's home state where there is not a serious risk that the claimant will be persecuted. If there is such a place, and it is not "unreasonable" to expect the specific claimant to go there, then the refugee claim will fail.

(31.) Report of the Secretary-General, In-depth study on all forms of violence against women, United Nations General Assembly (6 July 2006) A/61/122/Add.1 at para. 112.

(32.) Ibid. at para. 53.

(33.) Ibid. at paras. 52-56.

(34.) Cases reported in LexisNexis Quicklaw where reasons are dated from 2005 to 2009 (as of May 2010). See Appendix 1 for list of cases.

(35.) Report of the Secretary-General, supra note 31 at para. 113.

(36.) Canada (AG) v. Ward, [1993] 2 S.C.J. No 74 at para. 50.

(37.) Ibid. at para. 50.

(38.) Ibid. at para. 49.

(39.) Ibid. at para. 48.

(40.) YLF (Re), [2007] RPDD 148; BTF (Re), [2007] RPDD 88; Low v. Canada (MCI), [2006] RPDD 42; and Horvath v. Canada (MCI), [2004] RPDD 315.

(41.) Report of the Secretary-General, supra note 31 at para. 278

(42.) Ibid. at para. 78.

(43.) Ibid. at para. 295.

(44.) Ibid. at para. 281.

(45.) M.G.H. (Re) [2006] RPDD 184.

(46.) For example see Fontenelle v. Canada (MCI), [2006] RPDD 49; Lewis v. Canada (MCI), [2006] RPDD 48; Chaves v. Canada (MCI), [2004] RPDD 604; Ste-Marie v. Canada (MCI), [2004] RPDD 358; and Peters v. Canada, [2004] RPDD 556.

(47.) Lewis v. Canada, [2006] RPDD No 48.

(48.) Ibid. at para. 17.

(49.) Ibid. at para. 19.

(50.) [1990] 1 SCR852.

(51.) YLF,, supra note 40 at para. 12.

(52.) Shi v. Canada (MCI), [2004] RPDD 659 at para. 16.

(53.) BKD (Re), [2007] RPDD 197 at para. 17.

(54.) Williams v. Canada (MCI), [2004] RPDD 674 at para. 8.

(55.) HOZ v. Canada (MCI), [2004] RPDD 22 at para. 27.

(56.) This is a direct quote from the RPD reasons as cited in a judicial review of that decision in MDCR v. Canda, [2009] FCJ No. 370 at para. 21 (the RPD decision was released in 2008, but not reported in LexisNexis Quicklaw).

(57.) Hoz, supra note 55 at para. 15

(58.) VHF (Re), [2006] RPDD 158 at paras. 11 &12.

(59.) Dunsmuir v. New Brunswick [2008] 1 SCR 190 at para. 47, Canada (Citizenship and Immigration) v. Khosa 2009 SCC 12 at para. 46.

(60.) CIC, "Refugees: Refugee Claims in Canada--Refugee Appeal Division" online: www.cic.gc.ca/english/refugees/inside/appeals.asp (accessed June 13, 2010).

(61.) Mendoza v. Canada [2008] FCJ 481 at para. 15.

(62.) Alexander v. Canada, [2009] FCJ No. 1682 [emphasis in original].

(63.) See for example Meneses v. Canada [2009] FCJ No 846, M.D.C.R. v. Canada [2009] FCJ 370, Zepeda v. Canada, [2009] 1 F.C.R. 237, Huerta v. Canada, [2008] FCJ No. 737, Farias v. Canada, [2008] FCJ No. 1292, Low v. Canada [2007] FCJ No. 326, Hooper v. Canada [2007]FCJ No. 1744.

(64.) Jones v. Canada (MCI), [2006] FCJ No 591.

(65.) See for example Richie v. Canada, [2006] FCJ No. 210., Henry v. Canada [2006] FCJ 1341, Jones v. Canada [2006] FCJ No. 591, El Hage v. Canada [2008] FCJ No. 1459, Isakova v. Canada [2008] FCI No 188.

Constance MacIntosh is an associate professor at the Schulich School of Law, Dalhousie University, where she teaches in the area of immigration and refugee law. She is a proud graduate of Osgoode Hall Law School's Intensive Programme in Immigration and Refugee Law, and was recently appointed the Domain Leader for the Justice, Policing and Security Domain of the Metropolis Project (Atlantic Region).

Research support for this study was provided by the Trustees of the National Association of Women and the Law Charitable Trust for Research and Education, as a Policy Research Grants Project. Special thanks to Margaret Denike and Heather Neuman for their comments upon this report, to Yoko Yoshida for her assistance with translating data into charts and other forms, and to Brian Noble for his endless support. A full copy of this report, including a consolidated list of recommendations and appendixes that list all surveyed cases, can be found at http://www.nawl.ca.
Although gender is not specifically enumerated as one of the
   grounds for establishing Convention refugee status, the definition
   of Convention refugee may properly be interpreted as providing
   protection for women who demonstrate a well-founded fear of
   gender-related persecution by reason of any one, or a combination
   of, the enumerated grounds. (Emphasis added.)


... when considering whether it is objectively unreasonable for the
   claimant not to have sought the protection of the state, the
   decision maker should consider, among other relevant factors, the
   social, cultural, religious and economic context in which the
   claimant finds herself.


Implementation of international standards at the national level
   remains inadequate, as evidenced by the continuing prevalence of
   violence against women worldwide. Human rights treaty bodies
   regularly draw attention to gaps in the implementation of
   international obligations relating to violence against women. The
   Committee on the Elimination of Discrimination against women has
   consistently identified a number of areas of concern, as have other
   human rights treaty bodies. (41)


Essentially the claimant cites social and cultural norms as the
   reason why she had repeatedly failed to seek police assistance. In
   this instance, the Board cannot find fault with the security forces
   or legal system when the claimant was unwilling to avail herself of
   these services. Based on the information before the Board, then the
   objective basis of the claim fails. (48)


[The police] cannot offer help if they are unaware of the
   transgressions. The historical situation in Grenada for abused
   females suggests the system failed for two reasons: 1) police and
   legal disinterest, and 2) a failure on the part of victims to
   report abuse, due to their dependency on males for economic
   survival. (49)


the Board's curt dismissal of the Applicant's explanation [for
   remaining in the home] suggests that is [sic] was inconceivable
   that the Applicant would remain in an abusive relationship for any
   good reason and the fact that she did so must mean she is not
   telling the truth. The Supreme Court addressed implications such as
   these in relation to women suffering ongoing abuse in R. v.
   Lavallee, [1990] 1 S.C.R. 852, which is cited in the Gender
   Guidelines; in that case, this kind of implication was described as
   part of the "mythology about domestic violence." This is another
   clear example of how the Board failed to apply the Gender
   Guidelines in a meaningful way.


Chart 1: Overall acceptance rates by the Refugee
Protection Division for claims based on domestic
violence (2004-2009) *

Accepted claims    2%
Declined claims   98%

* Decisions reported in LexisNexis Quicklaw, 2004-2009 (as of
April 2009).

Note: Table made from pie chart.

Chart 2: RPD decisions on domestic violence
based claims (2004-2009)

Year       Reported        Accepted         Declined
           decisions *

2009          2             1                1
2008          9             0                9
2007          21            0                21
2006          23            0                23
2005          30            0                30
2004          50            2 ([dagger])     48
Total         135           3                132

* Claims reported in LexisNexis Quicklaw, 2004-2009 (as of
April 2010).

([dagger]) In [2004] RPDD No. 131, the adjudicator found there was no
objectively founded future risk of persecution. However, the
claim was accepted under the "compelling reasons" exception
(IRPA s.108(4)).

Chart 3: Grounds for declining claims in 132 RPD
decisions (2004-2009) *

Credibility (of
story of
domestic
violence)              46

Presumption
of state
protection
not rebutted           86

Lack of Nexus
with
enumerated
ground                  0

Internal Flight
Alternative            15

Delay
resulting in
finding of no
subjective             16

Identity not
established             1

Domestic
violence not
persecution             0

* Some claims were denied on multiple grounds. In those cases,
all grounds for denying the claim were recorded on this chart,
and so the columns do not add up to 132. Data reflects RPD
claims reported in LexisNexis Quicklaw, 2004-2009 (as of April
2010).

Note: Table made from bar graph.

Chart 4: RPD references to international human
rights instruments (2004-2009) *

Makes no reference to human
rights instruments                 82

Refers to state having signed a
human rights treaty                 4

* Data reflects claims reported in LexisNexis Quicklaw 2004-2009
(as of Apri1 2010) where the claim was rejected on grounds
including the presumption of state protection not being rebutted.
(N=86.)

Note: Table made from pie chart.

Chart 5: Observations on legislation that
criminalizes or makes domestic
violence actionable *

Observes law           3
does not exist

Does not discuss if   12
laws exist

Observes laws         29
exist and
comments on
effectiveness

Just observes that    42
laws exist

* This chart represents RPD decisions in cases where the presumption
of state protection was not rebutted. (N= 86.) Cases
reported on LexisNexis Quicklaw, 2004-2009.

Note: Table made from bar graph.

Chart 6: RPD Observations on police
responsiveness to women seeking protection
from domestic violence (2004-2009) * (N=86)

Observed          11
police
'reluctant' to
act

Observed           8
police were
usually
responsive

Observed          12
attitudes were
'improving'
(but no
objective
assessment)

Observed data      5
was
contradictory

No comment
on policing
responsiveness    50

* RPD decisions reported in LexisNexis Quicklaw, 2004-2009 (as
of Apri1 2010), where the claim was declined on grounds which
included the presumption of state protection not being rebutted.

Note: Table made from bar graph.

Chart 7: Socio-cultural contextual factors observed
by RPD adjudicators where state presumption of
ability to protect was not rebutted (2004-2009) *
(N=86)
                          Yes    no comment

Observed that them         17        69
are cultural/societal
attitudes of violence
against women

Observed that               6        80
domestic violence is
considered 'private'

Observed whether           51        35
domestic violence is
a serious/widespread
problem

Noted reports on           36        50
police responsiveness

Observed whether           21        65
women were
reluctant or did not
report domestic
violence

* Cases reported in LexisNexis Quicklaw, 2004-2009 (as of April
2010).

Note: Table made from bar graph.

Chart 8: RPD observations
on whether shelters exist *

observes    44%
whether
shelters
exist

does not    66%
consider
whether
shelters
exist

Note: Table made from pie chart.

Chart 9: If RPD observed shelters exist,
observations on adequacy *

finds are        0%
adequate
shelters

finds           34%
shelters are
inadequate
or do not
exist

no comment      66%
on adequacy

* These charts reflect RPD decisions where the presumption that
the state could protect was not rebutted. Decisions reported in
LexisNexis Quicklaw, 2004-2009 (as of April 2010).

Note: Table made from bar graph.

Chart 10: RPD observations on gender and
economic dependency* (N=86)

Observed women were         6%
generally economically
dependent upon men (6%)

Observed the claimant       1%
to be personally
dependent (1%)

No comment on gender       93%
and economic
dependency (93%)

* Chart reflects RPD decisions that were reported in LexisNexis
Quicklaw (2004-2009), as of Apri12010, where one ground on
which the claim was dismissed was that the presumption of state
protection was not rebutted.

Note: Table made from pie chart.

Chart 11: Judicial review rates of Negative RPD
decisions where claimant
was fleeing domestic violence (2005-2009)

Judicial review allowed    44%
RPD decision upheld        56%

* Cases reported in LexisNexis Quicklaw, 2005-2009 (N=89).

Note: Table made from pie chart.

Chart 12: Grounds for permitting judicial review *
(2005-2009)

Finding on state     77%
protection not
reasonable

Finding on           31%
credibility not
reasonable

Gender Guideline     33%
not followed

Evidence not          5%
considered

Error in FA           8%
analysis

Natural
justice/procedure     8%
fairness issue

* Cases reported in LexisNexis Quicklaw, 2005-2009 (N=39). In
some cases the application was granted on multiple grounds and
so the columns do not add up to 100 per cent.

Note: Table made from bar graph.
Gale Copyright: Copyright 2009 Gale, Cengage Learning. All rights reserved.