Date: 20141114

Docket: IMM-7980-13

Citation: 2014 FC 1080

[ENGLISH CERTIFIED TRANSLATION]

Montréal, Quebec, November 14, 2014

PRESENT: The Honourable Mr. Justice Martineau

BETWEEN:

MONIA PATRICIA DJOSSOU

Applicant

and

THE MINISTER OF CITIZENSHIP AND IMMIGRATION

Respondent

JUDGMENT AND REASONS

[1]               The applicant is challenging the legality of a decision of the Refugee Appeal Division [RAD] of the Immigration and Refugee Board [IRB] upholding a previous decision by the Refugee Protection Division [RPD] finding that the applicant was neither a “Convention refugee” nor a “person in need of protection” within the meaning of sections 96 and 97 of the Immigration and Refugee Protection Act, SC 2001, c 27 [IRPA or Act].

[2]               The applicant is a citizen of Benin who had claimed refugee protection following the persecution she purports to have suffered in Togo after the death of her Togolese husband. Her in-laws want to force her into a polygamous marriage with her brother-in-law; this is in addition to a rape and harassment she was subject to in that country. The applicant does not want to seek refuge in Benin, because she claims her family would force her to return to live with her brother-in-law, given that her father had already received payment for the dowry from her in-laws.

[3]               The RPD refused the refugee protection claim based on issues of credibility and because it determined the applicant’s conduct to be inconsistent with that of a person who alleges a fear of being persecuted in their country. In her appeal before the RAD, the applicant contended that the RPD erred in fact and in law: (1) in its assessment of the applicant’s credibility, by failing to take into account all of the evidence in the record; (2) in the manner in which if justified its negative determination, as insufficient reasons were provided for the decision.

[4]               The applicant asked the RAD to hold an oral hearing, but her request was denied by Member Bissonnette, who found that no new admissible evidence, meeting the requirements of subsection 110(4) of the Act, had been presented to the RAD. In this case, the applicant’s appeal was dismissed on the basis of the evidence in the record because the RPD’s decision “falls within a range of possible, acceptable outcomes which are defensible in respect of the facts and law”, while the RPD’s reasons for dismissing the claim “were sufficiently justified, transparent and intelligible”, hence this application for judicial review.

Systemic approach

[5]               Neither the interpretation of the RAD’s authority to admit new evidence nor the Member’s refusal to hold an oral hearing are at issue here (for an interesting study of the matter, I would recommend reading the judgment issued by Justice Gagné in Singh v Canada (Minister of Citizenship and Immigration), 2014 FC 1022 [Singh]). Thus, the only issue in this matter is determining whether the RAD committed a reviewable error in applying, to the RPD’s findings of fact or of mixed fact and law, the standard of “reasonableness” used by courts sitting in judicial review (Dunsmuir v New Brunswick, 2008 SCC 9 [Dunsmuir]). One may, in this regard, speak of a systemic approach on the part of the RAD.

[6]               Indeed, this is not the first time the legality of decisions founded on the same legal reasoning as that employed by this and other members of the RAD has been examined on judicial review by judges of this Court. I refer you to the judgments below:

1.      Iyamuremye v Canada (Citizenship and Immigration), 2014 FC 494 [Iyamuremye] (Justice Shore), setting aside a decision dated July 25, 2013, by Member Bissonnette (Docket IMM-5282-13);

2.      Triastcin v Canada (Citizenship and Immigration), 2014 FC 975 [Triastcin] (Justice Shore), setting aside a decision dated August 26, 2013, by Member Bissonnette (Docket IMM-5981-13);

3.      Akuffo v Canada (Citizenship and Immigration), 2014 FC 1063 [Akuffo] (Justice Gagné), upholding a decision dated September 18, 2013, by Member Gallagher (Docket IMM-6640-13);

4.      Alvarez v Canada (Citizenship and Immigration), 2014 FC 702 [Alvarez] (Justice Shore), setting aside a decision dated October 18, 2013, by Member Bissonnette (Docket IMM‑7218-13);

5.      Eng v Canada (Citizenship and Immigration), 2014 FC 711 [Eng] (Justice Shore), setting aside a decision dated October 22, 2013, by Member Bissonnette (Docket IMM‑7281-13);

6.      Njeukam v Canada (Citizenship and Immigration), 2014 FC 859 [Njeukam] (Justice Locke), upholding a decision dated October 22, 2013, by Member Bissonnette (Docket IMM-7280-13);

7.      Yetna v Canada (Citizenship and Immigration), 2014 FC 858 [Yetna] (Justice Locke), setting aside a decision dated November 5, 2013, by Member Leduc (Docket IMM‑7567‑13);

8.      Spasoja v Canada (Citizenship and Immigration), 2014 FC 913 [Spasoja] (Justice Roy), setting aside a decision dated November 8, 2013, Member Bissonnette (Docket IMM‑7630-13);

9.      Huruglica v Canada (Citizenship and Immigration), 2014 FC 799 [Huruglica] (Justice Phelan), setting aside a decision dated September 5, 2013, by Member Bosveld (IMM‑6362-13);

10.  Diarra v Canada (Citizenship and Immigration), 2014 FC 1009 [Diarra] (Justice Beaudry), setting aside a decision dated January 23, 2014, by Member Leduc (Docket IMM-1217-14);

11.  Guardado v Canada (Citizenship and Immigration), 2014 FC 953 [Guardado] (Justice Martineau), setting aside a decision dated 24, 2014, by Member Bissonnette (Docket IMM-882-14);

12.  Alyafi v Canada (Citizenship and Immigration), 2014 FC 952 [Alyafi] (Justice Martineau), setting aside a decision dated January 30, 2014, by Member Gallagher (Docket IMM-1091-14).

[7]               As can be seen from a review of the reasons provided by the Court in the various matters noted above, a number of colleagues and I are of the unanimous view that the RAD is committing a reviewable error when it adopts a reasonableness standard of review, even if there are varying opinions as to the nature or scope of an appeal before the RAD. All of these RAD decisions share the same common characteristic; they were all issued between July 24, 2013, and January 30, 2014, some four months before this Court was first asked to examine the legality of a RAD decision in which this same reasoning was applied (Iyamuremye, above, dated May 26, 2014). In all but two instances in which it was determined that the results were reasonable despite faulty reasoning by the Member (Njeukam and Akuffo, above), the Court decided to set aside the RAD’s decision and return the appeal to the RAD for redetermination.

General position of the parties

[8]               Unsurprisingly, the applicant is arguing that the decision dated November 7, 2013, by Member Bissonnette should meet the same fate as the other RAD decisions set aside by the Court. In the present case, correctness is the applicable standard of review. Indeed, sections 110 and 111 of the IRPA make to reference to any standard of review or to any concept of deference whatsoever. If the decision of the RPD is wrong in law, in fact or in mixed law and fact, the RAD may set it aside and substitute the determination that should have been made. Parliament in fact intended to create a “full appeal” before the RAD. In this case, it was not enough for Member Bissonnette to simply examine the reasonableness of the RPD decision to reject the claim and inquire only whether it fell “within a range of possible, acceptable outcomes which are defensible in respect of the facts and law” (Dunsmuir, above at para 47). In particular, to determine whether the RPD had erred in its assessment of the applicant’s credibility—by disregarding explanations or documents produced— the RAD would necessarily have to reassess the evidence as a whole and reach its own conclusions as to the merits of the claim.

[9]               Reprising arguments that so far have yet to be accepted by this Court, the respondent reiterates that, on the contrary, Member Bissonnette committed no reviewable error by not reassessing all of the evidence in the record and by applying a standard of review of reasonableness; that a reasonableness standard should be applied to the review of the decision by the RAD. At any rate, the applicant’s proposition that the RAD must proceed with an “independent review of the evidence” and which is supported by a certain amount of the case law of this Court is not consistent in law and disregards the scheme of the new statutory provisions. The wording of section 110 of the IRPA requires a refugee protection claimant to identify any error of fact, of law or of mixed law and fact, made by the RPD at first instance. There is no mention of having a trial de novo or of reassessing all of the evidence in the record. Lastly, the respondent adds that even if the standard of review to be applied to the RPD’s findings of fact or of mixed fact and law was that of palpable and overriding error, as certain colleagues of this Court assert, the end result should be the same, which would justify the dismissal of this application for judicial review.

[10]           At the hearing, in response to the Court’s questions regarding the case law, the applicant’s learned counsel argued that if the Court would not decide, the issue of the scope of the appeal before the RAD on a correctness standard, it could at least include, in its reasons for judgment, some clarification with regard to the scope of sections 110 and 111 of the IRPA. The Court’s opinion could be of great assistance to the parties and to the panel when the issue comes up again for redetermination, given that the statements in the case law with respect to the applicable standard can be obscure and contradictory at times. In particular, counsel for the applicant questioned the scope of the test proposed by Justice Phelan in Huruglica, above, according to which “the RAD is required to conduct a hybrid appeal”; on the one hand, the Court indicates that the RAD “must review all aspects of the RPD’s decision and come to an independent assessment” (at para 54) while on the other, the Court states that the RAD “can recognize and respect the conclusion of the RPD on such issues as credibility and/or where the RPD enjoys a particular advantage in reaching such a conclusion” (at para 55). These statements appear to be contradictory. The applicant’s counsel further opined that the automatic application of the palpable and overriding error standard, which is what Justices Shore and Roy appear to favour (Alvares, Eng and Spasoja, above) is likely to lead to serious problems of application and create injustices in the future because one cannot compare the RAD to a traditional court of appeal that hears thousands of different cases.

[11]           At first glance, both parties agree that it its current wording the standard of palpable and overriding error set out by the Supreme Court in Housen v Nikolaisen, 2002 SCC 33 [Housen] and HL v Canada (Attorney General), 2005 SCC 25 [HL] seems more demanding than the reasonableness standard set out in Dunsmuir, above. It should be recalled that in the latter case the Supreme Court decided to meld the two previous standards (reasonableness simpliciter and patent unreasonableness) into a single standard of review. If a greater degree of deference (patent and overriding error) poses no problem for the respondent, it makes no practical sense for the applicant. The applicant points out that under the new scheme thousands of refugee claimants rejected by the RPD are now denied an opportunity to have a pre-removal risk assessment [PRRA]. And cases that are appealed before the RAD are already restricted to certain countries in which peoples’ lives are often at stake. For the applicant, these latter factors make a strong case for the RAD adopting a standard of review for RPD decisions that affords precious little room for deference, and one that is certainly less deferential than the reasonableness standard, no matter which qualifiers are used (reasonableness or palpable and overriding error).

[12]           For the reasons that follow, this application for judicial review should be allowed.

Standard of review

[13]           Let us begin this analysis by determining which standard of review this Court, as a superior court sitting in judicial review of any decision made under the IRPA (sections 3, 18 and 18.1 of the Federal Courts Act, RSC 1985, c F-7; section 72 of the IRPA), must apply to various determinations of the RAD (Dunsmuir at paras 27 et seq.). Any deference the RAD may or may not give to the RPD raises a question of law. There are two competing standards: correctness and reasonableness.

[14]           It is well known in legal circles that the application of a correctness standard by a court on judicial review to a question of law determined by an administrative tribunal is more stringent than applying a standard of reasonableness. Indeed, “[u]nlike a review for correctness, there will often be no single right answer to the questions that are under review against the standard of reasonableness” (Law Society of New Brunswick v Ryan, 2003 SCC 20 at para 51). In contrast, on a correctness standard there can only be one right answer. It goes without saying that in certain cases this could be determinative of the outcome of an application for judicial review (as was the case, for example, in Singh, above at para 65; see also King v Canada (Attorney General), 2012 FC 488 at paras 94, 144-145 [King], affirmed by 2013 FCA 131).

[15]           But why speak of deference?

[16]           It is because, as a general rule, the more deferential standard of reasonableness will usually apply where a decision maker is interpreting its own statute or statutes closely connected to its function, with which it will have particular familiarity (Dunsmuir, above at paras 54 and 55; Smith v Alliance Pipeline Ltd, 2011 SCC 7, [2011] 1 S.C.R. 160 at para 28; Alberta (Information and Privacy Commissioner) v Alberta Teachers’ Association, 2011 SCC 61, [2011] 3 S.C.R. 654 at para 30 [Alberta Teachers’ Association]; Canadian Artists’ Representation v National Gallery of Canada, 2014 SCC 42 at para 13). As the Supreme Court of Canada aptly summarizes in Canadian National Railway Co v Canada (Attorney General), 2014 SCC 40 at para 55 [Canadian National Railway Co.]:

In such cases, there is a presumption of deferential review, unless the question at issue falls into one of the categories to which the correctness standard applies: constitutional questions, questions of law that are of central importance to the legal system as a whole and that are outside of the adjudicator’s expertise, questions regarding the jurisdictional lines between two or more competing specialized tribunals, and the exceptional category of true questions of jurisdiction (Dunsmuir, at paras. 58-61, and Alberta Teachers’ Association, at para. 30, citing Canada (Canadian Human Rights Commission), at para. 18, and Dunsmuir, at paras. 58-61).

[17]           In Dunsmuir, above, the Supreme Court notes that when a full analysis of the applicable standard of review must be carried out, the analysis must be contextual, and it must take into consideration the factors relevant to the determination of the applicable standard (at para 64):

As mentioned above, it is dependent on the application of a number of relevant factors, including: (1) the presence or absence of a privative clause; (2) the purpose of the tribunal as determined by interpretation of enabling legislation; (3) the nature of the question at issue, and; (4) the expertise of the tribunal.  In many cases, it will not be necessary to consider all of the factors, as some of them may be determinative in the application of the reasonableness standard in a specific case.

[18]           One must begin by asking whether the level of deference to be accorded to a particular type of question has been “established satisfactorily” in the case law (Agraira v Canada (Public Safety and Emergency Preparedness), 2013 SCC 36 at paras 48-49). Despite the fact that some of my colleagues have heretofore opted for a correctness standard (Iyamuremye, above at para 20; Alvarez, above at para 17; Eng, above at para 18; Huruglica, above at paras 25-34; Yetna, above at para 14; Spasoja, above at paras 7 to 9), there is nothing approaching unanimity on the issue (contra, Akuffo, above at paras 16 to 26). I myself am of the view, although I may be mistaken, that a reasonableness standard applies to this review of the legality of the decision in question.

[19]           In the first place, it is by no means evident to me that the issue here falls under one of the categories of questions that are subject to a correctness standard (Dunsmuir, above at paras 58-61; Canadian National Railway Co., above at para 55). To begin with, in this case, no constitutional question has been raised by the parties. In addition, I would doubt that any veritable “question of jurisdiction” in its narrow sense is at play here, which is the case “where the tribunal must explicitly determine whether its statutory grant of power gives it the authority to decide a particular matter” (Dunsmuir, above at para 59). It should be recalled that the Supreme Court of Canada warned that federal courts “should not be alert to brand as jurisdictional, and therefore subject to broader curial review, that which may be doubtfully so” (Dunsmuir, above at para 35, citing Canadian Union of Public Employees, Local 963 v New Brunswick Liquor Corp.,, [1979] 2 SCR 227 at p. 233).

[20]           According to subsections 110(1) and (2) of the IRPA, an appeal before the RAD involves “a question of law, of fact or mixed law and fact” raised by the refugee claimant or by the Minister in his or her notice of appeal. In this case, the parties agree that the RAD was properly seized with the applicant’s appeal. The RAD has the express power to overturn any RPD decision that is “wrong in law, in fact, or in mixed law and fact” (paragraph 111(2)(a) of the IRPA). The scope of the appeal is therefore not in issue. Nor is it a matter of interpreting the scope of the exclusions set out in subsection 110(2)—which restrict the RAD’s capacity to hear certain types of appeal. Rather, the issue is to determine whether the member committed a reviewable error in choosing to apply a reasonableness standard when considering questions of fact or of mixed law and fact raised by the applicant in her notice of appeal (Dunsmuir, above), and did so on the basis of a questionable interpretation or application of the judgment issued by the Alberta Court of Appeal in Newton v Criminal Trial Lawyers’ Association, 2010 ABCA 399, 493 AR 89 [Newton].

[21]           Indeed, for Member Bissonnette, only a pure question of law or a breach of natural justice are reviewable on a correctness standard on appeal before the RAD (impugned decision at para 38). Otherwise, the appellant has the onus of demonstrating to the RAD the “unreasonableness” of the RPD’s findings of fact or of mixed law and fact that she is challenging (impugned decision at paras 39 to 41). But can one consider the issue of “standards of review” before the RAD as being a “question of jurisdiction”, in the broad sense, because it would be incidental to the delineation of the “respective jurisdictions” of the RAD and the RPD?

[22]           At least this is what my colleague, Justice Roy, suggests in Spasoja (above at para 8). With respect, I am not convinced that we need to go down that road. Indeed, unlike an adjudicator or a human rights tribunal—which can both be seized with, at first instance, an issue related to employment discrimination—the RPD and the RAD do not have competing jurisdiction in determining refugee status—other than where the RAD, on appeal, decides to set aside the decision of the RPD and substitute the decision that should have been made, without referring the matter back to the RPD, as is permitted under section 111 of the IRPA.

[23]           Questions of statutory interpretation are indubitably questions of law (Canadian National Railway Co., above at para 33). Such is therefore the case where the RAD is interpreting its enabling statute, in this case the IRPA. In practice, deference is largely an incidental issue to the RAD’s perception of its appellate role. In the absence of an explicit statutory provision, one might say it is a question of “judicial or institutional policy”—for lack of a better description. It must be noted here that neither section 110 nor section 111 of the IRPA make any specific reference to the “degree of deference” the RAD may or may not afford to a finding of fact, of law, or of mixed law and fact made by the RPD. In passing, the concept of deference that we associate with “standards of review” should not be confused with the particular grounds for appeal or for review of an appealable or reviewable decision (Alyafi, above at paras 14-15). Therefore, if it is not a true question of jurisdiction, does the issue fall under another category to which a correctness standard applies?

[24]           Up to this point, the Supreme Court has given a very narrow reading of the last exception, that of a question of law of central importance. In fact it was on the basis of this narrow reading that my colleague, Justice Gagné, recently concluded that a standard of reasonableness should be applied when determining which standard the RAD must apply when reviewing findings of fact made by the RPD (Akuffo, above at paras 17-26). Justice Gagné noted that a correctness standard would apply solely to questions of law that were both of central importance to the legal system as a whole and outside the expertise of the administrative tribunal (at para 20). Citing Alberta Teachers’ Association, above, Justice Gagné indicated that a question of general importance is one whose resolution has repercussions outside the statutory scheme under consideration (Akuffo, above at para 21). Moreover, since its decision in Alberta Teachers, above, the Supreme Court has not encountered any situation that would fall under this exception to the reasonableness standard (Akuffo, above at para 21), since questions whose resolution would have no precedential value outside of a specific context are not questions of central importance to the legal system as a whole.

[25]           In Nor-Man Regional Health Authority Inc v Manitoba Association of Health Care Professionals, 2011 SCC 59 [Nor-Man] and Communications, Energy and Paperworkers Union of Canada, Local 30 v Irving Pulp & Paper, Ltd., 2013 SCC 34 [Irving], the Court reaffirmed that arbitration decisions issued under collective agreements were subject to a reasonableness standard, even if the arbitrator applied the equitable doctrine of estoppel (Nor-Man, above at para 38) and even if the conflict was of interest to the public (Irving, above at para 66). In Irving, Justices Rothstein and Moldaver, for the minority (dissenting on another issue), pointed out that even if the dispute was of wider importance to the public, the applicable standard was reasonableness because the application of collective agreements is part of labour arbitrators’ expertise and that “[t]his dispute has little legal consequence outside the sphere of labour law and that, not its potential real-world consequences, determines the applicable standard of review” (at para 66). The Supreme Court arrived at a similar conclusion in Canadian National Railway Co., above, in which it pointed out the issue as to whether certain parties could avail themselves of the complaint mechanism under the Transportation Act, SC 1996, c 10, was not a true question of jurisdiction or a question of central importance because the question at issue does not have any precedential value outside of its statutory regime (at paras 60-62).

[26]           In McLean v British Columbia (Securities Commission), 2013 SCC 67, the Supreme Court reaffirmed the presumption that a reasonableness standard should be applied to an administrative decision maker’s interpretation of its home statute (at para 21). Justice Moldaver, on behalf of the majority, noted that an administrative decision maker’s interpretation of a limitation period contained in its enabling statute did not automatically attract a standard of correctness:

First, although I agree that limitation periods, as a conceptual matter, are generally of central importance to the fair administration of justice, it does not follow that the Commission’s interpretation of this limitation period must be reviewed for its correctness. (at para 28)

[27]           The Supreme Court also refuted the appellant’s argument that limitation periods were not part of the substantial securities regulation in which the Commission had a specialized expertise (at para 30). According to the Court:

. . . [T]he resolution of unclear language in an administrative decision maker’s home statute is usually best left to the decision maker.  That is so because the choice between multiple reasonable interpretations will often involve policy considerations that we presume the legislature desired the administrative decision maker — not the courts — to make.  Indeed, the exercise of that interpretative discretion is part of an administrative decision maker’s “expertise”. (at para 33)

[28]           This conclusion is similar to that found in Canada (Human Rights Commission) v Canada (Attorney General), 2011 SCC 53 [Canada (Human Rights Commission)], wherein the Supreme Court indicated that the standard of reasonableness applied to the Canadian Human Rights Commission’s decision that it could award costs under its enabling statute because this was a question of law that was within the core expertise of the Tribunal in the interpretation and application of its enabling statute (at para 25). The Court further stated:

In addition, a decision as to whether a particular tribunal will grant a particular type of compensation — in this case, legal costs — can hardly be said to be a question of central importance for the Canadian legal system and outside the specialized expertise of the adjudicator. Compensation is frequently awarded in various circumstances and under many schemes. It cannot be said that a decision on whether to grant legal costs as an element of that compensation and about their amount would subvert the legal system, even if a reviewing court found it to be in error. (at para 25)

[29]           The jurisprudence of the Supreme Court shows that where a response to a question has no precedential value outside of a particular statutory scheme, it is not a question of central importance to the legal system as a whole. These decisions also demonstrate that a broad interpretation must be given to an administrative tribunal’s expertise in applying its enabling statute or a statute closely connected to its functions, in particular, that a tribunal’s expertise is not limited to substantive provisions, but extends to procedural provisions.

[30]           With respect, I am not as convinced as my colleague, Justice Phelan, that “[t]he selection of the appropriate standard of review is a legal question well beyond the scope of the RAD’s expertise, even though it depends on the interpretation of the IRPA, the RAD’s home statute” (Huruglica, above at para 30). At first glance, given the experience of its members and institutional expertise, the RAD is very well placed to determine whether the new statutory provisions create a “true appeal”, an “appeal de novo”, or another type of administrative appeal. Moreover, the RAD’s need for expert members was one of the reasons cited by M.P. Nina Grewal to explain why the government was not in favour of establishing the RAD in 2007:

It should also be noted that in order to implement the RAD, the IRB itself has said that the skill set of members of the RAD would need to be different from other IRB members. The IRB stated that the selection would have to reflect the tasks of an appellate decision-maker, require a stronger legal and analytical capacity, and some prior adjudicative experience. (House of Commons Debates, 39th Parliament, 1st Session, No. 122 (March 2, 2007) at pp. 1330 et seq.))

[31]           Furthermore, comments by Peter Showler, Chairperson of the IRB at the time, before the Standing Committee on Citizenship and Immigration during its 2001 review of Bill C-11, which went on to become the IRPA, show that the purpose for creating the RAD was not to simply add an intermediate stage between the RPD’s determination of a refugee claim and judicial review by the Federal Court:

It is expected that the RAD will produce two different but complementary results. By reviewing individual RPD decisions on the merits, the RAD can efficiently remedy errors made by the RPD. That, if you will, is the safety net for the RPD. However, in addition the divisions will ensure consistency in refugee decision-making by developing coherent national jurisprudence in refugee law issues. As I said to this committee before, we don't see that as a benefit simply in that it will improve the quality of our decision-making. If there is more coherent, consistent jurisprudence, we think RPD decision-makers can actually make their decisions more quickly as well. (Standing Committee on Citizenship and Immigration, Evidence, 37th Parliament, 1st Session, meeting No. 5 (March 20, 2001) at pp. 0915-20, 0945; Emphasis added.)

[32]           In addition, the IRPA sets out the following at subsection 162(1):

[33]           It is therefore apparent that RAD has jurisdiction over any question of law that is presented to it, including the standard of review it should apply. The RAD’s specialization, and the expertise of its members, as demonstrated by its function of standardization of law and the precedential value of decisions of three members pursuant to paragraph 171(c) of the IRPA, indicates that the Federal Court must defer to the RAD. Further, although there are differences in the manner in which the Immigration Appeal Division [IAD] and the RAD hear or decide appeals that may be before them, their respective decisions are protected by the same privative clause (section 162 of the IRPA); the members of both divisions have considerable expertise in determining appeals under the IRPA; and both divisions have the authority to render the decision that should have been rendered by the original decision maker. Or, in the case of the IAD, the Supreme Court decided that, taken together, those factors clearly point to the application of a reasonableness standard of review to decisions issued under section 67 of the IRPA (Canada (Citizenship and Immigration) v Khosa, 2009 SCC 12, [2009] 1 S.C.R. 339 at para 58 [Khosa]). Why would it be otherwise for RAD decisions issued under section 111 of the IRPA?

[34]           Accordingly, unless one is convinced that “[s]etting the standard of review [applicable to a specialized appellate tribunal’s review of a lower administrative tribunal’s decision] is a legitimate aspect of the superior court’s supervisory role” (Newton, above at para 39, cited in Huruglica, above at para 27), and that one considers, moreover, that it is a question of law “of central  importance for the legal system as a whole”, the Court should now avoid unilaterally proclaiming which standards of deference are to be applied to RPD decisions in an appeal before the RAD.

[35]           Nevertheless, the Court in Huruglica, above, did certify the following question of law: “What is the scope of the Refugee Appeal Division’s review when considering an appeal of a decision of the Refugee Protection Division?” The respondent has since filed a notice of appeal with the Federal Court of Appeal [A-470-14].

[36]           A Federal Court judge does not have the luxury of being wrong about the applicable standard of review for RAD decisions. At this stage, one cannot therefore assume that the Federal Court of Appeal—if it agrees to answer a question as general as that of the Court—will answer the question certified by Justice Phelan by reviewing the RAD’s decision on a correctness standard. For example, in B010 v Canada (Citizenship and Immigration), 2013 FCA 87, the Federal Court of Appeal applied a reasonableness standard to the Immigration Division’s interpretation of paragraph 37(1)(b) of the IRPA and responded to the certified question by stating that “it is reasonable to define inadmissibility under paragraph 37(1)(b) by relying upon subsection 117(1) of the Immigration and Refugee Protection Act …”. If the Federal Court of Appeal finds, in Huruglica, that a reasonableness standard applies instead, it could rephrase the question certified by Justice Phelan in such a manner so as to ask whether the option selected by the RAD (the reasonableness approach borrowed from Dunsmuir, above) was an acceptable outcome in respect of sections 110 and 111 of the IRPA. Moreover, it could also choose to respond instead to the questions that were certified by Justice Gagné (Akuffo, above at para 53), in the event an applicant whose claim for refugee protection has been rejected were to appeal a negative Federal Court judgment to the Federal Court of Appeal.

[37]           Hence, this is why it seemed to me to be more prudent, in this case, as long as the issue has not been finally resolved on appeal, to adopt a pragmatic approach. Also, for the very reasons that were put forth in Alyafi, above, I do not think there is any need, at this particular moment, for me to make any sort of final ruling on the interpretation of sections 110 and 111 of the IRPA to determine this application for judicial review. For the time being, there appear to be a number of possible approaches, but what is clear, however, is that the option chosen by the RAD (a judicial review-based approach) is not an acceptable outcome in law. Even applying the lesser standard of reasonableness, I still arrive at the same end result as my colleagues who applied the more stringent correctness standard. Intervention is warranted in this case. In this way, the choice of appropriate standard of review will not be determinative of the matter (which might not have been the case had I adopted a correctness standard or had I dismissed the applicant’s application by applying a standard of reasonableness).

An appeal is not a judicial review

[38]           First, it is important to remember that the reasonableness standard should not be seen as a plenary dispensation for decisions of expert decision-makers. Even if an interpretation of the law made by a specialized tribunal has to be reviewed on a reasonableness standard, it remains that the interpretation of the law is always contextual. The law does not operate in a vacuum and the tribunal is always required to take into account the legal context in which it is called to apply the law (see King, above at para 60; Dunsmuir, above at para 74).

[39]           The fundamental problem in this case is that the legal reasoning by Member Bissonnette (see paragraphs 30 to 31 of the decision under review), appears to me to be unreasonable on its face because the RAD cannot, in practice, conduct itself in an appeal as a judicial court sitting in judicial review (Alyafi, above at paras 10-18 and 53; Spasoja, above at paras 3, 9, 11 and 47; Huruglica, above at paras 39-54). Otherwise, the creation of a specialized appeal tribunal for refugee determination would serve no purpose (Alyafi, above para 12).

[40]           In the present case, subsections 110(1) and 111(1) and (2) of the IRPA state:

[41]           It is clear from reading the aforementioned provisions that the RAD can set aside the RPD’s decision and substitute the decision that, in its opinion, should have been made, which means that the RAD has much broader powers on appeal that those of a traditional court of law sitting in judicial review. Not only that, the RAD may, among other things, admit new evidence and decide to hold an oral hearing in specific circumstances set out by Parliament (subsections 110(3) to (6) of the IRPA). Further, the RAD exercises exclusive jurisdiction on appeal that is at least equal to that of the RPD at first instance (subsection 162(1) of the IRPA) and can itself render the decision that ought to have been rendered by the RPD (section 111 of the IRPA). Such is not the case with the Federal Court, whose jurisdiction is limited by sections 72 to 75 of the IRPA, as well as by sections 18 and 18.1 of the Federal Courts Act. In addition, the remedies available to the Federal Court are limited in principle to setting aside the decision and remitting the matter for redetermination, which is not the case with the RAD vis-à-vis the RPD.

[42]           Reasonableness is a well-known standard in judicial review, which follows a different line of reasoning than that of an appeal (Alyafi, above at paras 17 and 18). A reviewing court applying this standard must ask itself whether the decision under review and its justification possess the “attributes” of reasonableness. In practice, reasonableness is concerned mostly with the existence of justification, transparency and intelligibility within the decision-making process. But it is also concerned with whether the decision falls within a range of possible, acceptable outcomes which are defensible in respect of the facts and law (Dunsmuir, above at para 47). But why would an appellate body adopt such an approach when it was created for the express purpose of hearing appeals – thus adding another level of adjudication – and when its decisions are themselves subject to judicial review?

[43]           I share the view expressed on this subject by my colleague Justice Phelan in Huruglica, above at paras 39, 41-43:

[39] In considering the nature of the review to be conducted by the RAD, if the RAD simply reviews RPD decisions for reasonableness, then its appellate role is curtailed. It would merely duplicate what occurs on a judicial review. Further, if the RAD only performed a duplicative role to that of the Federal Court, it would be inconsistent with the creation of the RAD and the extensive legislative framework of the IRPA.

 …

[41] In legal terms, the creation of an appellate tribunal would suggest that Parliament sought to achieve something other than that available under judicial review. In the British Columbia Court of Appeal decision of British Columbia Society for the Prevention of Cruelty to Animals v British Columbia (Farm Industry Review Board), 2013 BCSC 2331, 237 ACWS (3d) 16 [BC SPCA], the matter under review was the creation of an internal appeal between the first level decision and judicial review. The Court held that the appeal was to be substantive.

[42] In BC SPCA, at paragraph 40, that court summarized the above principle which is equally applicable in the present case:

Logically, if the legislature had intended the deferential sort of review for which the SPCA contends, it would have amended nothing and left the whole matter to the process of judicial review. That, however, was what the legislature hoped to avoid. To do so, it created a brand-new appeal process to the FIRB. The result, surely, was not meant to be just a different venue for the same process as before.

[43] It flows that in creating an internal appellate body, within the executive branch of government, the principle of standard of review, a function of the division of powers between the executive and the judiciary, is of lesser importance and applicability. The traditional standard of review analysis is not required.

[44]           In this case, the parties raised no new argument, and provided no particular reason in this case, that would allow me to distinguish the aforementioned decisions or to depart from the legal reasoning by which an appeal is not a judicial review, and which has, to date, been adopted by the Court. I must therefore conclude that the applicant was denied the right to an appeal under the Act; this constitutes a reviewable and determinative error according to the near-unanimous case law of this Court.

But what type of appeal might this be?

[45]           A critical flaw in the decision under review—and in other RAD decisions set aside by the Court—is that before establishing any sort of standard of review, it is imperative that the following question be answered: What type of appeal are we speaking of?

[46]           There is general agreement that there are usually three types of appeal: true appeal (“appel veritable”); appeal de novo; and hybrid appeal. Frank Falzon provides the following overview:

3. There are three general types of appeals to specialized administrative tribunals. The most narrow is what Dupras [v Mason, 1994 CanLII 2772 (BC CA)] refers to as a true appeal, where the appeal is founded on the record and where the appellant must demonstrate a reviewable error of law, fact or procedure. The broadest is what Dupras describes as an appeal de novo, where the original decision is ignored in all respects, except possibly for purposes of cross-examination. The third is a mixed model of appeal in which the appellant retains the onus of demonstrating error and the appeal board receives the record, but the appeal is not limited as to grounds, the appeal board reviews the decision below for correctness and fresh evidence may be adduced without constraint. These three broad models are conceptual starting points, and are subject to variation according to the specific intent of the governing legislation. Appeals to Administrative Tribunals (2005) 18 Can J Admin L & Prac 1 at pp. 34-35.

[47]           The lax use of the terms “appeal de novo”, “appeal”, or “full appeal” can only add to the confusion that seems to exist among parties and attorneys. In this regard, from a legal perspective, what distinguishes an appeal de novo from a true appeal is that in an appeal de novo, the matter is heard as if it was at first instance: the second decision maker is not required to identify an error of fact or of law made by the initial decision maker (Dupras v Mason, 1994 CanLII 2772 (BC CA)). In short, the decision under appeal is owed no deference. In that sense, an appeal before the RAD therefore resembles, at first glance, a true appeal, but it may also be a hybrid appeal. Indeed, if certain colleagues of mine express the view that an appeal before the RAD is perhaps not an appeal de novo in the strict sense of the term, they do not exclude the possibility of reweighing the evidence that was before the RAD (Iyamuremye, above at para 35; Eng, above at para 26; Alvarez, above at para 25; Huruglica, above at paras 52 and 54).

[48]           It should be noted that a statutory text may specify that an appeal is heard de novo, but this is not always the case. Regard must be had in particular to the legislative context of the nature of the bodies in question and the impact of the decisions on individuals’ rights. For example, section 63 of the IRPA (former sections 79 and 77 of the Immigration Act, RSC 1985, c I-2, since repealed) does not expressly provide that the IAD may hear an appeal de novo. Nonetheless, according to the case law, appeals from an immigration officer’s refusal to issue a permanent resident visa to a sponsored member of the family class are heard de novo by the IAD (Mohamed v Canada (Minister of Employment and Immigration), [1986] 3 FCR 90 at paras 9-13; Kahlon v Canada (Minister of Employment and Immigration), 14 ACWS (3d) 81, [1989] FCJ No 104 (CAF) at para 5; Kwan v Canada (Minister of Citizenship and Immigration), 2001 FCT 971, [2001] FCJ No 1333 at paras 15-18 [Kwan]).

[49]           As Justice Muldoon notes in Kwan at para 17:

[17] The Court does not accept the applicant's contention that a hearing de novo applies uniquely to errors of fact. A hearing de novo is undertaken as if the matter were before the Appeal Division for the first time, and the issue is not how the visa officer came to her conclusion, but whether the sponsoree is a member of the family class. An appeal under subsection 77(3) is not a judicial review, but an entirely new hearing in which the Board examines the whole record and hears submissions by the appellant and a case officer.

[50]           The third category encompasses so-called “hybrid” appeals, which may include a more traditional review of decisions in order to verify whether any errors in fact, law, or mixed fact and law were made by the initial decision maker, as well as a de novo review of the matter by the second decision maker. For example, such is the case with appeals of discretionary orders of Federal Court prothonotaries (Canada c Aqua-Gem Investments Ltd, [1993] 2 RCF 425, 1993 CanLII 2939 (CAF) [Aqua-Gem]; Merck & Co, Inc c Apotex Inc, 2003 CAF 488 at paras 17-28). In that regard, it is interesting to note that where an order issued by a prothonotary raises questions that are vital to the final issue of the case, a Federal Court judge must exercise his or her own discretion de novo, therefore, by conducting a hearing de novo, even if no new evidence has been adduced by the appellant. Otherwise, as in a true appeal, the judge merely considers whether the prothonotary was “clearly wrong, in the sense that the exercise of discretion by the prothonotary was based upon a wrong principle or upon a misapprehension of the facts” (Aqua-Gem, above).

[51]           One may also speak of a hybrid model in the area of trademarks. In the case of an appeal from a Registrar’s decision, section 56 of the Trade-marks Act, RSC 1985, c T-13, provides that the Federal Court may consider new evidence; it is not solely an appeal based on the record that was before the Registrar. Nonetheless, “some deference” is afforded to the administrative decision maker. In Molson Breweries v John Labatt Ltd., [2000] FCJ No 159, [2000] 3 FCR 145 (FCA), the Federal Court of Appeal explained that it was neither a customary appeal, nor an appeal de novo in the strict sense:

[46] Because of the opportunity to adduce additional evidence, section 56 is not a customary appeal provision in which an appellate court decides the appeal on the basis of the record before the court whose decision is being appealed. A customary appeal is not precluded if no additional evidence is adduced, but it is not restricted in that manner. Nor is the appeal a "trial de novo" in the strict sense of that term. The normal use of that term is in reference to a trial in which an entirely new record is created, as if there had been no trial in the first instance.¹² Indeed, in a trial de novo, the case is to be decided only on the new record and without regard to the evidence adduced in prior proceedings.¹³

[^[12 Black's Law Dictionary, 7th ed. (St. Paul, Minn.: West Group, 1999) defines a "trial de novo" as: "A new trial on the entire case" that is, on both questions of fact and issues of law" conducted as if there had been no trial in the first instance."]

[47]      On an appeal under section 56, the record created before the Registrar forms the basis of the evidence before the Trial Division judge hearing the appeal, which evidence may be added to by the parties. Thus, although the term trial de novo has come into frequent usage in describing a section 56 appeal, the term is not an entirely accurate description of the nature of such an appeal. That an appeal under section 56is not a trial de novo in the strict sense of the term was noted by McNair J. in Philip Morris Inc. v. Imperial Tobacco Ltd. (No. 1).¹⁴

[48]      An appeal under section 56 involves, at least in part, a review of the findings of the Registrar. In conducting that review, because expertise on the part of the Registrar is recognized, decisions of the Registrar are entitled to some deference.

[52]           In this case, there was no meaningful analysis by Member Bissonnette of the nature of the appeal before the RAD. His conclusion as to the process the RAD must follow to hear an appeal is, with all due respect, unreasonable. The Member ought to have done more than review the RPD’s decision on the basis of the nature of the issue criterion that is more often than not automatically applied by courts sitting in judicial review. As a specialized administrative appeal tribunal, the RAD should now ask whether the appeal process provided at sections 110 and 111 of the IRPA, is a true appeal, an appeal de novo, or a hybrid appeal. If so-called “paper-based” appeals are the rule, and some parallel can reasonably be drawn with a true appeal (not a judicial review), the RAD may also, in the exercise of its discretion, consider new documentary evidence adduced by the refugee protection claimant or by the Minister and hold an oral hearing to hear viva voce evidence where the conditions set out at subsections 110(3) to (6) of the IRPA are met, in its view.

[53]           Although my colleague Justice Roy dismissed any suggestion that an appeal before the RAD is [translation] “an opportunity for a new trial or a reconsideration of the matter in its entirety” (Spasoja, above at para 39), other colleagues of mine, Justices Shore, Phelan and Gagné are not as categorical and all three insist on the need for a re-examination of the evidence even in paper-based appeals (Alvarez, above at paras 25 and 33; Eng, above at paras 26 and 34; Huruglica, above at paras 47, 48 and 52; Akuffo, above at para 45). Without deciding in favour of either approach, it is precisely this kind of reflection and analysis of possible options that is sorely lacking in the decision under review, thus rendering it unreasonable.

[54]           In this regard, in an article entitled “Refugee Appeal Division (RAD)—First Steps in an Important Legal Evolution” (2014) Imm L R (4th) 169, Mario Bellissimo and Joanna Mennie, specialized practitioners, argue that a “one size fits all” approach to RAD appeals is not consistent with the statutory framework. They indicate that, where RAD members have experience and skills that are superior to those of RPD members, the RAD should not show significant deference to the RPD. In addition, in order for the RAD to play a significant role and not be a mere intermediary between the RPD and judicial review in Federal Court, the RAD must not be overly deferential to the RPD’s findings. Furthermore, the RPD is in no better position than the RAD to assess viva voce evidence when the RAD convenes a hearing, which favours a nuanced approach that allows a different standard to be applied to different cases:

In a sense, the RAD could be described as a form of hybrid: it has the formal written argument, structured timelines, and quashing powers of the Federal Court, yet it also has the power to advance its own decision central to the concept of a de novo appeal.

[55]           Without making a final determination on the issue for the moment, the RAD should consider the three options (true appeal; appeal de novo; hybrid appeal) with an open mind.

Choosing an intervention model consistent with the wording and purpose of the Act

[56]           While recognizing that an error of law committed by the RPD is reviewable on a correctness standard (hence, without any deference), Member Bissonnette adopted a deferential approach with respect to the RPD’s findings of fact or mixed law and fact. But under what logic or legal principle?

[57]           Member Bissonnette’s reasoning is based first on the premise that “the mere presence of a right to appeal—including appeals within an administrative structure—in no way means that no deference to the first-level decision-maker is called for” (impugned decision at para 33 and note 28). That is at least what the Alberta Court of Appeal seems to indicate in Newton; it saw no objection in principle to extending to the administrative sphere (this was an appeal in relation to a police officer’s conduct) the standards of deference established by the Supreme Court in judicial review cases (Khosa, above; Dunsmuir, above) or traditional appeals (Housen, above; HL, above).

[58]           Member Bissonnette then introduced practical considerations from the perspective of the negative perception the public could have about the appeal process if the RAD conducted an independent assessment of the evidence in the record, hence “the importance of promoting the autonomy of the proceeding and its integrity” (impugned decision at para 35) and also of recognizing “the expertise and advantageous position of the first-level decision-maker” (impugned decision at para 36).

[59]           In particular, Member Bissonnette considered that he had the same kind of relationship vis-à-vis a member of the RPD as “a trial judge and an appeal judge” (impugned decision at para 34), and he referred to the following passage from Housen (above at para 17), which the Alberta Court of Appeal also cited in Newton (above at para 81):

The presumption underlying the structure of our court system is that a trial judge is competent to decide the case before him or her, and that a just and fair outcome will result from the trial process.  Frequent and unlimited appeals would undermine this presumption and weaken public confidence in the trial process.  An appeal is the exception rather than the rule.

[60]           Member Bissonnette’s reasoning appears unreasonable to me in this case.

[61]           The first step is always to read the legislation carefully. Indeed, it may sometimes specify the appropriate standard of review. For example, paragraph 18.1(4)(d) of the Federal Courts Act allows the Federal Court to intervene on judicial review when it is satisfied that the federal board, commission or other tribunal “based its decision or order on an erroneous finding of fact that it made in a perverse or capricious manner or without regard for the material before it”. In Khosa, above, the Supreme Court of Canada held that, in enacting paragraph 18.1(4)d), “Parliament intended administrative fact finding to command a high degree of deference” and that “[this paragraph] provides legislative precision to the reasonableness standard of review of factual issues in cases falling under the Federal Courts Act” (at para 46). In this case, which deals with an appeal before the RAD, sections 110 and 111 of the IRPA do not contain any particular qualifier and refer only to a decision of the RPD that “is wrong in law, in fact or in mixed law and fact” (see paragraph 111(2)(a) of the IRPA).

[62]           Through interpretation, one can read in a statute words that are not in the statute for the purpose of expanding—“reading in” or restricting “reading down”—the scope of the words used by the legislator. Such exercises are known in constitutional cases to “save” a statutory provision. We think, for example, of what the Supreme Court did in R v Sharpe, [2001] 1 SCR 45, 2001 SCC 2, by adding to the Criminal Code exceptions to the prohibition on possessing pornographic material. That being said, “reading in” is prohibited in statutory interpretation. As Professor Ruth Sullivan summarized it in Sullivan on the Construction of Statutes, 5th edition, (LexisNexis, 2008) at page 168: “ . . .  reading in may on occasion be justified as a constitutional remedy, it is not a legitimate interpretation technique. It amounts to amendment rather then paraphrase.”

[63]           Given that one of the cardinal rules of interpretation is that where a statute is clear, it is not necessary for the decision‑maker to discover the legislative intent, it is sufficient to apply it (R v Multiform Manufacturing Co, [1990] 2 S.C.R. 624, 1990 CanLII 79 (SCC); R v Clarke, 2014 SCC 28 at paras 11-12). Prima facie, when one reads sections 110 and 111 of the IRPA, one reaches the following conclusion: any error of fact, law or mixed law and fact committed by the RPD justifies the intervention of the RAD and the substitution of the impugned determination by the determination that, in its opinion, the RPD should have made in the case. There is no question in the statute of any “deference”, and I am not ready to accept, today, without a convincing legal demonstration, that there is a universal presumption of deference to findings of fact or of mixed law and fact made by decision‑makers at first instance. If Parliament had intended that a determination could only be set aside on appeal because an error of law was committed, why did it bother to add errors of fact or mixed law and fact?

[64]           In the absence of a specific legislative indication, the “degree of deference” that the RAD owes to a determination by the RPD flows from either a common law rule (Khosa, above, at paras 4, 18, 19, 26, and 42 to 51) or from the exercise of administrative discretion. Given that the RAD is not a court and does not exercise any superintending power vis-à-vis the RPD, I would opt for the second alternative. Indeed, all Canadian courts are bound by common law rules, but they are not bound by the degree of deference the RAD applies to determinations by the RPD. While the decisions made by three members of the RAD under section 171(c) of the IRPA have precedential value for panels of one member of the RAD and the RPD, they have no precedential value for courts, including the Federal Court.

[65]           I note that Member Bissonnette did not question whether the practical considerations that led traditional courts to adopt a deferential attitude so that “[a]n appeal is the exception rather than the rule” have the same weight in determining refugee status. In the context of refugees—persons who by definition are extremely vulnerable—the objectives of the Act are to recognize that the refugee program is in the first instance about saving lives and offering protection to the displaced and persecuted; to grant, as a fundamental expression of Canada’s humanitarian ideals, fair consideration to those who come to Canada claiming persecution; and to establish fair and efficient procedures that will maintain the integrity of the Canadian refugee protection system, while upholding Canada’s respect for the human rights and fundamental freedoms of all human beings (paragraphs 3(2)(a), (c) and (e) of the IRPA).

[66]           What, then, must the RPD and the RAD do to ensure the integrity of the Canadian refugee protection system and, here, why establish an appeal process if the RAD is to act like a traditional court sitting on judicial review or on appeal?

[67]           That is the question that until now has been avoided by the RAD, which has chosen to adopt an appeal intervention model directly copied from that of traditional appellate courts. In many respects, the refugee determination system is unique. First, the evidentiary rules are quite different from those that a judge at first instance applies in a civil or criminal matter. A refugee claimant who testifies at a hearing before the RPD does not have to repeat before the member the allegations in the narrative accompanying the refugee application form. The member plays a much more active role than an ordinary judge.

[68]           Under the IRPA, the members of the RPD—and also the members of the RAD—have the powers and authority of a commissioner appointed under the Inquiries Act (section 165 IRPA). They may inquire into any matter that they consider relevant to establishing whether a claim is well‑founded (section 170(a) of the IRPA). In other words, even before the hearing takes place, the RPD member will have already identified the issues that must be resolved, and it is normally the member who begins questioning the refugee claimant. As the Chairperson Guidelines 7 properly notes, “[a] member’s role is different from the role of a judge. A judge’s primary role is to consider the evidence and arguments that the opposing parties choose to present; it is not to tell parties how to present their cases. . . . The members have to be actively involved to make the RPD’s inquiry process work properly.”

[69]           On the other hand, apart from the inquiry process—inquisitorial in many respects—surrounding the particular circumstances of the claim, the IRB’s specialized staff prepares and updates what are called, in the vernacular jargon of practitioners, “national documentation packages” (NDPs). The onus is on the persons participating in refugee proceedings to consult the IRB’s website to review the documents in the NDP on the refugee claimant’s country of origin because the RPD could examine them in the context of a refugee claim in order to issue a decision. Moreover, the RPD may decide to use other documents as well, for example, reports produced by the IRB’s Research Directorate, media articles or reports from human rights organizations.

[70]           From the perspective of establishing facts, determining whether there is a well‑founded fear of persecution requires assessing a refugee claimant’s subjective fear—regarding not only the credibility of his or her narrative—but also its objective basis in light of the documentary evidence pertaining to the conditions in the country in question. On appeal, the RAD will also have access to the RPD’s record (including recordings) and all the documentary evidence (including the NDP of the country in question). Apart from a pure credibility issue (in passing, what is credibility?), one may reasonably ask whether a RAD member is in just as good a position as a RPD member to reassess the evidence in the record where it is alleged on appeal that the RPD erred in its assessment of the evidence as a whole, which is precisely the principal complaint that the applicant made against the RPD. A number of my colleagues think so, and I am also of that opinion.

[71]           Alas, in his analysis of the “standards of review”, it seems that Member Bissonnette did not find it useful to conduct an extensive analysis of the RAD’s new role, the wording of the new statutory provisions or the history that led to the enactment and coming into force—delayed by the government for a long time—of sections 110 and 111 of the IRPA, which were amended on a number of occasions since the IRPA was enacted in 2001 (see the Balanced Refugee Reform Act, SC 2010 c 8; the Protecting Canada’s Immigration System Act, SC 2012, c 17). There is also no question in the decision under review of the similarities and differences with the other statutory appeal scheme at the IRB, the one administered by the IAD—which would have been very useful in establishing the degree of deference that should be shown to the RPD’s findings of fact, law and mixed law and fact.

[72]           I have serious reservations about the exportation into the administrative sphere, even more so in refugee matters, of general criteria that were developed by traditional appellate courts. That was the case in Newton. It is incumbent on the RAD to develop its own criteria. Because a word of caution is necessary. The Court previously decided that the RAD committed a “reviewable error” by applying judicial review standards to RPD decisions and that the RAD must not merely duplicate the judicial review power of this Court (Huruglica, above at para 39; Spasoja, above at paras 21-24; Alyafi, above at paras 13-18). Note, however, in order to avoid [translation] “a blurring of lines” (Spasoja, above at para 21), I do not believe that the RAD must automatically apply judicial appeal standards (British Columbia Chicken Marketing Board v British Columbia Marketing Board, 2002 BCCA 473 at paras 13-14; Paul v British Columbia (Forest Appeals Commission), 2003 SCC 55 at para 44; and Whitehorse (City) v Yukon (YTCA), [1988] YJ No 5, 52 DLR (4th) 749). This naturally includes the reading that could be done of the Court’s decisions where the issue is applying the palpable and overriding error standard.

[73]           Moreover, although the decision under review briefly alludes to the factors enumerated by the Alberta Court of Appeal (Newton, above at para 43), the member’s analysis gives short shrift—except perhaps obliquely and superficially—to the other factors that are relevant in the particular context (Huruglica, above, at para 20) and to the importance and weight that must be assigned to each factor, having regard to the objectives stated in subsections 3(2) and (3) of the Act, the wording of sections 110 and 111 and the overall scheme of the Act. It seems necessary to us to look back in time.

Choosing an appeal model that also responds to the legitimate expectations of those most directly affected

[74]           At one time, ordinary federal public servants decided the merits of refugee claims without a hearing, while the Immigration Act, 1976, SC 1976-77, c 52, allowed the Immigration Appeal Board to redetermine a claim. After the Supreme Court of Canada decision in Singh v Minister of Employment and Immigration, [1985] 1 S.C.R. 177, which determined that the former system was inconsistent with the requirements of fundamental justice, the RPD came into existence under the name Convention Refugee Determination Division [CRDD]. Until the enactment of the IRPA, the CRDD was governed by sections 67-69.3 of the Immigration and Refugee Act, RSC 1985, c I-2. Refugee claims were decided by a quorum of the CRDD composed of two members—unless the person consented to his or her case being determined by a single member (subs 69.1(7) and (8) of the Immigration and Refugee Act). There was no appeal from an unfavourable decision. The only recourse was judicial review.

[75]           When the IRPA was enacted in 2001, the CRDD became the RPD. After the IRPA came into force on June 28, 2002, the RPD began hearing all cases, with some exceptions, before a single member (section 163 of the IRPA). At the same time, the IRPA contained sections 110 and 111 establishing the RAD as a specialized refugee appeal tribunal. The right of appeal was not limited except where the claim was determined to be abandoned or withdrawn. All appeals were decided on the record without a hearing. Despite this, the coming into force of sections 110 and 111 (amended twice since their enactment in 2001) was delayed until the RAD was established on December 15, 2012.

[76]           Hansard remains a very useful contextual tool to identify the purpose of a statute and the reasons why Parliament decided to intervene, although the reliability and weight of Parliamentary debates are limited (Canadian National Railway Co. v. Canada, above at para 47). Moreover, even though great care must be taken in assigning weight to them, proposed but unenacted provisions may also provide information about the purpose of the legislation and give an indication of legislative intent (Canada (Canadian Human Rights Commission), above at para 44).

[77]           When Bill C-11, now the IRPA, was tabled, all of the stakeholders took the position that the introduction of RPD panels composed of a single member was offset by the introduction of a right of appeal (see, inter alia, the testimony of Joan Atkinson, Assistant Deputy Minister, Operations and Program Management, Citizenship and Immigration Canada, before the Standing Committee on Citizenship and Immigration, 37th Parliament, 1st Session, meeting No. 27 (May 17, 2001) at p 1140). In addition, Minister Caplan, who was responsible for the bill, made it clear at the Senate hearings that, in establishing the RAD, “[t]he whole purpose is to ensure that the correct decision is made and that people are given an appeal” (Standing Senate Committee on Social Affairs, Science and Technology, Evidence, 37th Parliament, 1st Session, Issue 29 (October 4, 2001)).

[78]           Peter Showler, former Chairperson of the IRB, also expressed an opinion on reducing RPD hearings from two members to one:

In contrast to the present model, where claims are normally heard by two-member panels, the vast majority of protection decisions will be made by a single member. Single-member panels are a far more efficient means of determining claims. It is true that claimants will no longer enjoy the benefit of the doubt currently accorded them with two-member panels, and I think that should be noted. However, any perceived disadvantage is more than offset by the creation of the refugee appeal division, the RAD, where all refused claimants and the minister have a right of appeal on RPD decisions.

Appeals to the RAD will be in writing only and will be reviewed by experienced RPD decision-makers with the power to affirm the RPD decision, to set it aside and substitute their own decision, or to refer the matter back to the RPD for a rehearing on particular issues in exceptional cases where it might be necessary to hear additional evidence. We estimate the workload of the RAD will be about 8,000 to 9,000 cases per year, and we intend to equip the division with a corresponding level of staff and resources. (Standing Committee on Citizenship and Immigration, Evidence, 37th Parliament, 1st Session, meeting No. 5 (March 20, 2001) at pp. 0915-20, 0945; Emphasis added.)

[79]           Let me open another equally important parenthesis. We are in 2007. Already more than five years have passed since the IRPA came into force in 2002, and the RAD has still not been established, which is of great concern to parliamentarians. MP Nicole Demers introduced Bill C‑280, An Act to Amend the Immigration and Refugee Protection Act (Coming into Force of Sections 110, 111 and 171), a bill that would, in the end, not be enacted. At second reading of the bill, MP Richard Nadeau referred to a number of systemic considerations justifying the establishment of the RAD raised by François Crépeau, then an international law professor at the University of Montréal and now the United Nations’ Special Rapporteur on the Human Rights of Migrants and professor at McGill University (see House of Commons Debates, 39th Parliament, 1st Session, No. 122 (March 2, 2007) at pp 1400 et seq).

[80]           These considerations, also cited by the Canadian Council for Refugees, were efficiency, consistency of the law, justice and reputation:

The Refugee Appeal Division is indispensable for the smooth functioning of the Canadian refugee determination system for four reasons:

• The first reason is efficiency. A specialized appeal division for refugee matters can deal much more efficiently with unsuccessful claimants than the Federal Court, an application for pre-removal risk assessment or requests on humanitarian grounds. The refugee appeals division can do a better job of correcting errors of law and fact.

• The second reason is consistency of the law. An appeal division deciding on the merits of the case is the only body able to ensure consistency of jurisprudence both in the analysis of facts and in the interpretations of legal concepts in the largest administrative tribunal in Canada.

• The third reason has to do with justice. The decision to refuse refugee status has extremely serious consequences, including death, torture, detention, and so on. As in matters of criminal law, the right to appeal to a higher tribunal is essential for the proper administration of justice.

• [translation of first five words only] “The fourth reason is reputation: as a procedural safeguard, the Refugee Appeal Division will enhance the credibility of the IRB in the eyes of the general public, just as the provincial Courts of Appeal reinforce the entire justice system. The IRB's detractors--both those who call it too lax, and those who call it too strict--will have far fewer opportunities to back up their criticisms and the Canadian refugee determination system will be better able to defend its reputation for high quality.

Parliament did not disregard these considerations, and they still apply today where the issue is defining the scope of the appeal before the RAD and the important role it plays in the refugee determination process.

[81]           In 2010, a new bill was tabled in the House, Bill C-11, the Balanced Refugee Reform Act. The amendments to sections 110 and 111 of the IRPA are important because they expand the scope of appeals before the RAD by providing the opportunity to present new evidence and obtain an oral hearing before the RAD. That being said, the basic principles outlined above remain the same as in 2001: a right of appeal on any question of fact, law or mixed law and fact, and the RAD’s power to confirm the determination, substitute a determination that, in its opinion, should have been made for the one made by the RPD or refer the matter to the RPD.

[82]           Before the Standing Committee on Citizenship and Immigration, the Minister of Citizenship and Immigration, the Honourable Jason Kenney, commented on the establishment of a RAD with more powers than the one contemplated in 2001:

However, there is finally an appeal section, which is even better than what was provided by the legislation in 2002.

This new appeal division would provide most claimants with a second chance, an opportunity to introduce new evidence about their claim and to do so in an oral hearing, if necessary. And, significantly, Mr. Chairman, the bill would make it possible to remove those who would abuse our system within a year of their final IRB decision.

I want to underscore that the refugee appeal division foreseen in the Immigration and Refugee Protection Act 2003, and proposed, for instance, in Mr. St-Cyr's private member's bill, does not actually include, as does the RAD in Bill C-11, the ability to present new evidence and in certain cases to have an oral hearing before the appeal division decision-maker. This is an improved RAD. It's an additional level of administrative fairness, but it's not going to happen if we don't achieve the other streamlining in the system that the package speaks to.

(Standing Committee on Citizenship and Immigration, Evidence, 40th Parliament, 3rd Session, meeting No. 12 (May 4, 2010) at pp. 1535, 1610; Emphasis added.)

[83]           Before the Senate Committee, Minister Kenney added:

The result would be a streamlined system that would actually add greater procedural fairness, through the creation of what's known as the Refugee Appeal Division. This would allow failed claimants a full appeal of their claims.

In terms of our system, Bill C-11 would provide for the following. First, the creation of a new interview with an Immigration and Refugee Board public servant, in place of a written form, early in the claims process. In our opinion, that would speed up the process and make it more efficient. Second, independent decision makers at the Refugee Protection Division of the IRB who are public servants rather than political appointees. That means that people who hold the hearings for asylum claimants will be, after those reforms, IRB officials rather than cabinet appointees. Third, a new fact-based refugee appeal division that even surpasses what refugee advocates have requested for a long time.

…

The initial hearing at the Refugee Protection Division and the appeal at the Refugee Appeal Division both constitute an analysis of the risk faced by the claimant. Will they face a risk of torture or threat to their life if returned to their country of origin? . . . Our position is that once you have had two negative risk assessments — that is, once an IRB officer has looked at your case and said that you do not face risk if returned to your country and a refugee appeal decision maker has made the same decision — we do not think it is appropriate to have a third, redundant, risk assessment based on that legal criteria of risk, which is now embedded in sections 96 and 97 of the Immigration and Refugee Protection Act.

(Standing Senate Committee on Social Affairs, Science and Technology, Evidence, 40th Parliament, 3rd Session, Issue 11 (June 22, 2010). Emphasis added.)

[84]           The amendments to sections 110 and 111 enacted by Parliament in 2010 would not, however, come into force. Moreover, with the new amendments to sections 110 and 111 of the IRPA enacted in 2012 by Bill C-31, in addition to the 2010 amendments, the issue now is reducing the cases where an appeal can be heard by the RAD through limiting the eligible countries. Another limit relevant to this case is that no appeal lies to the RAD from any RPD decision that states that the claim has no credible basis or is considered manifestly unfounded (paragraph 111(2)(c) of the IRPA). As the Federal Court of Appeal stated in Rahaman v Canada (Minister of Citizenship and Immigration), 2002 FCA 89, the threshold for a no credible basis finding is very high because the RPD is required to “examine all the evidence and to conclude that the claim has no credible basis only when there is no trustworthy or credible evidence that could support a recognition of the claim” (at para 51). The underpinning for a determination that a claim has no credible basis is therefore a finding on the applicant’s credibility made by the RPD. While section 111 of the IRPA provides that these determinations cannot be appealed to the RAD, section 111 does not restrict the right of appeal on other credibility findings made by the RPD.

[85]           Moreover, despite the new restrictions to section 111 set out in Bill C-31, the possibility of access to a fact-based appeal is always mentioned. In particular, Justice Phelan in Huruglica (above at para 40) referred to what Minister Kenney said on second reading with respect to Bill C-31 in March 2012:

I reiterate that the bill would also create the new refugee appeal division. The vast majority of claimants who are coming from countries that do normally produce refugees would for the first time, if rejected at the refugee protection division, have access to a full fact-based appeal at the refugee appeal division of the IRB. This is the first government to have created a full fact-based appeal. [Emphasis added.]

[86]           As the Court has already pointed out, “whether it [the objective of creating a true appeal] was achieved is another question” (Huruglica, above at para 40). For his part, in Spasoja (above at paras 32-38), Justice Roy also cites numerous excerpts of the same type. In light of the passages cited in his decision, it appears clear that the creation of the RAD had a dual purpose: (1) on the one hand, to enable the RAD to efficiently correct errors made by the RPD by conducting a complete review of questions of fact, law and mixed law and fact and (2) on the other hand, to enable the RAD to ensure consistency in the decision-making process by establishing uniform jurisprudence on refugee law issues. I agree completely with the opinion of my colleague Justice Roy that [translation] “nothing in 2010 would suggest that appeals would have the appearance of quasi-judicial review”, while [translation] “[t]he same generous appeal theme was addressed by the Minister when he introduced Bill C-31, which became the Protecting Canada’s Immigration System Act, SC 2012, c 17, for second reading in the House of Commons” (Spasoja, above at paras 36-37).

[87]           We paraphrase by saying that the RAD was created to ensure that the RPD makes [translation] “correct decisions” and that it applies [translation] “correct law”  to the facts of the case. But the interpretation proposed by Member Bissonnette in the impugned decision is far from a [translation] “generous appeal” or a [translation] “complete appeal”. The goal appears to be to discourage appeals on questions of fact or mixed law and fact by introducing a concept of deference directly copied from the judicial model. The RAD’s current reasoning fails to take into account the factors that are relevant in the particular context of the Act, one being [translation] “that the legislative scheme, viewed as a whole, does not at all suggest deference within the meaning of the reasonableness standard” and the other that [translation] “[j]udicial review, with its inherent deference, stems from a very different logic than an appeal” (Spasoja, above at paras 20-21).

Conclusion

[88]           I have concluded for the above-noted reasons that Member Bissonnette’s reasoning in this case is not an acceptable outcome in law. The respondent invites me, in spite of everything today, in the exercise of my discretion, to dismiss this application for review because there is a risk that the end result will be the same. We note that section 18.1 of the Federal Courts Act “generally sets out threshold grounds which permit but do not require the court to grant relief” (Khosa, above at para 36). In this regard, as the Supreme Court points out, “[w]hether or not the court should exercise its discretion in favour of the application will depend on the court’s appreciation of the respective roles of the courts and the administration as well as the ‘circumstances of each case’: see Harelkin v. University of Regina, [1974] 2 S.C.R. 561, p. 575” (Khosa, above at para 36).

[89]           On judicial review, the Court must primarily ensure that the process followed by the administrative tribunal is consistent with the Act and does not lead to a real or apprehended injustice. The RAD is a specialized appeal tribunal, and the Court is not constituted to reassess the evidence. Unlike the RAD, the Court does not sit on appeal from decisions of the RPD. The court is therefore not well positioned to substitute itself for the RAD to assess the merits of the parties’ arguments in favour of or against the RPD’s decision, and even less, to review the evidence in the tribunal record in light of a standard of deference that the RAD never applied before—even assuming that there should be some deference with respect to questions of fact or mixed law and fact.

[90]           On the other hand, I am not convinced in this case that the result of the applicant’s appeal will be the same if the RAD applies a different standard than the one it has applied until now (Dunsmuir, above). At this time, some of my colleagues appear to favour the application of the so-called palpable and overriding error standard (Alvarez, Eng and Spasoja, above) while others categorically reject this approach (Huruglica, above at para 55) or ignore it in practice (Njeukam and Yetna, above). The latter propose instead applying what I myself characterized as “a composite and variable standard of review” (Alyafi, above at para 16).

[91]           At this stage, until such time as the Federal Court of Appeal or even the Supreme Court of Canada decides the issue definitively, it does not appear appropriate to me to judicially impose any standard of deference (Alyafi, above at paras 51‑52) on the RAD. In the interim, unless the Court or the RAD orders a stay of proceedings, the onus will be on the RAD to review the Act and to adopt a new test based on an analysis that will, this time, take all the relevant factors into account. That being said, I do not believe that the two alternative approaches (the so-called “palpable and overriding error” appellate standard of review; a composite and variable standard of review) discussed in Alyafi, above, are the only options to consider—the absence of deference in the case of any error of law, fact or law and fact being also a possible option.

[92]           The application for judicial review will therefore be allowed, and the case will be referred back to the RAD for a reconsideration of the applicant’s appeal that takes into account the jurisprudence of the Court and these reasons. While the applicant had a question to propose if her application for judicial review was dismissed, the respondent did not have any question to propose to the Court.

JUDGMENT

THE COURT ORDERS AND ADJUDGES that the application for judicial review is allowed. The impugned decision is set aside, and the matter is referred back to the Refugee Appeal Division for a reconsideration of the applicant’s appeal that takes into account the jurisprudence of the Court and the reasons accompanying this judgment. No question is certified.

“Luc Martineau”

FEDERAL COURT

SOLICITORS OF RECORD

DOCKET:		IMM-7980-13
STYLE OF CAUSE:		MONIA PATRICIA DJOSSOU v MINISTER OF CITIZENSHIP AND IMMIGRATION
PLACE OF HEARING:	Montréal, QuEbec
DATE OF HEARING:	OCTOBER 22, 2014
JUDGMENT AND REASONS:	MARTINEAU J.
DATED:	NOVEMBER 14, 2014

APPEARANCES:

Stéphanie Valois	FOR THE APPLICANT
Suzanne Trudel	FOR THE RESPONDENT

SOLICITORS OF RECORD:

Stéphanie Valois Counsel Montréal, Quebec	FOR THE APPLICANT
William F. Pentney Deputy Attorney General of Canada Montréal, Quebec	FOR THE RESPONDENT