One of the questions the members of the RAD have been asking themselves in their decisions to date is whether they should be deferential to findings made by the RPD. My presentation focused on this. My answer was a heavily qualified yes: the RAD should show some deference to the RPD, but it should not apply rigidly tests developed for review by courts of administrative decisions.
Plainly some account must be taken of the decision to create two separate bodies. Their functions must be somewhat distinct. Moreover, while the RAD is largely a 'paper tribunal', the RPD sees witnesses in the flesh which, according to the Supreme Court of Canada, is an advantage in assessing credibility (see here, para. 25). Finally, the individual participates in a meaningful way before the RPD, but will only in limited circumstances appear before the RAD. Participation may lead to more accurate outcomes and, of course, gives the individual a greater sense of 'ownership' of the decision, thereby responding to her dignitary interests.
These considerations rule out an entirely de novo appeal. Yet in defining its role by reference to the relevant statutory provisions and the need to allocate decision-making authority between it and the RPD, multiple alternative possibilities are open to the RAD. It could act like a reviewing court. It could take a robust approach to correcting errors by the RPD which, in the RAD's view, compromise the integrity of the refugee system. It could simply give weight to the conclusions of the RPD and reserve final decisions to itself. Or it could treat the RPD's role as being primarily one of gathering the relevant facts, leaving the RAD with a broad mandate to make binding conclusions in individual cases and thereby enforce the objectives of the statute.
The statute gives the RAD the power to hear appeals on any issue of law, mixed law and fact or fact, which is a power conferred in the broadest terms (though it does not, I argued, require the RAD to reason in terms of the problematic law/fact distinction). It allows it to substitute its judgment for that of the RPD or to remit a matter for disposition in accordance with directions. And it provides that decisions of three-member panels have the force of precedent, pointing to a policy-development role for the RAD.
Another important consideration is that the RAD and RPD are both part of the same entity, the Immigration and Refugee Board. They are to work in harmony, in pursuit of the same goals.
It would thus be a mistake to import wholesale the approach to judicial review set out in Dunsmuir, Khosa, Newfoundland Nurses, Agraira and so on.
For example, it would be wrong for the RAD to ask simply whether a decision falls within a range of possible, acceptable outcomes (Dunsmuir). The statute indicates that it is for the RAD to determine what is possible and acceptable in terms of the legislation. Similarly, it would be entirely inappropriate for the RAD to refuse to intervene because an appellant is asking it to reweigh factors on the basis that the RPD gave disproportionate weight to one consideration (Khosa). Determining the appropriate weight to give to relevant factors is as much a task for the RAD as it is for the RPD.
Moreover, it is not enough, in my view, for the RAD to ask itself whether it can "clearly understand" (Agraira) the basis of decision. Rather, it should ask whether the reasons are convincing. To put it another way, are these reasons that are acceptable to the Immigration and Refugee Board and the individual(s) in question? Or still another way, can the Board confidently sign off on them? The problem that arose in Newfoundland Nurses, of an outside court setting unrealistically high standards for written reasons, does not arise here, because it is part of the RAD's role to set standards for the RPD.
Concerns about autonomy point towards less deferential review. Individuals aggrieved at the outcome of their application can make an application for judicial review in Federal Court. If judicial review before the Federal Court is provided for, that suggests that the RAD should not exercise a reviewing function. Otherwise it is simply duplicating the role of the Federal Court.
Indeed, if the Federal Court has to review the RAD for reasonableness, there will be a "double deference" problem: the RAD will have reviewed the RPD for reasonableness, so the Federal Court will have to ask itself whether the RAD's application of reasonableness was reasonable. In these circumstances, I think it is highly likely that the Federal Court (whether it admits it or not) would step into the shoes of the RAD, thereby appropriating its role.
All of these considerations point towards less deferential review. This could take several forms: (a) a much more relaxed application of the reasonableness standard applied by courts; (b) giving weight to the conclusions of the RPD but reserving final decisions to the RAD; or (c) treating the RPD as a fact-gathering body. Given the assumption that the RPD has an edge on credibility issues, (c) is probably out. But both (a) and (b) are live options and, in my view, there is much to commend (b). It ensures the RAD plays a policy-development role but also ensures that there is a role for the RPD, whose conclusions will be treated with respect.
The same conclusions largely apply to appeals on procedural matters. The RAD is a paper tribunal. The usual argument in favour of intrusive review on matters of procedure is that courts are expert because they deal with procedural matters all the time. This may not be true in the case of the RAD.
However, it is clear that where an internal administrative appeal exists, reviewing courts will generally insist that the individual takes the appeal first. In order to safeguard the integrity of the administrative decision-making system, the appellate body must perform a robust review of procedural errors, as was demanded by the Supreme Court of Canada in Harelkin v. University of Regina,  2 SCR 561:
Sections 78(1)(c) and 33 (1)(e) are in my view inspired by the general intent of the Legislature that intestine grievances preferably be resolved internally by the means provided in the Act, the university thus being given the chance to correct its own errors, consonantly with the traditional autonomy of universities as well as with expeditiousness and low cost for the public and the members of the university. While of course not amounting to privative clauses, provisions like ss. 55, 66, 33(1)(e) and 78(1)(c) are a clear signal to the courts that they should use restraint and be slow to intervene in university affairs by means of discretionary writs whenever it is still possible for the university to correct its errors with its own institutional means. (my emphasis and see also here for a review of the relevant principles)Plainly, on some matters, which amount more to good case management than to violations of procedural rights, deference may be appropriate. Yet, when it comes to the potential for deference on questions of procedural fairness, in light of Harelkin and the policy-making role I see attributed to the RAD, I doubt there is much scope for the RAD to take a deferential approach.