Certified Questions, References and Reasonableness: In Canadian immigration law, there is a ‘certified question’ regime.

In my post on “Unreasonable Bilingual Interpretations of Law“, I mentioned that the Supreme Court would have the opportunity in Canada (Citizenship and Immigration) v. Mason, 2021 FCA 156 (leave granted) to say more about the methodology of reviewing administrative interpretations of law.

Mason raises other issues as well, one of which also arose in Canada (Citizenship and Immigration) v. Galindo Camayo, 2022 FCA 50.

In Canadian immigration law, there is a ‘certified question’ regime. Decisions made by the Immigration and Refugee Board are subject to judicial review in the Federal Court but “an appeal to the Federal Court of Appeal may be made only if, in rendering judgment, the judge certifies that a serious question of general importance is involved and states the question” (Immigration and Refugee Protection Act, SC 2001, c 27, s. 74(d)).

Prior to Vavilov, it was settled law (1) that a question may be certified by the Federal Court only where it is “a serious question that is dispositive of the appeal, transcends the interests of the parties and raises an issue of broad significance or general importance”: Lunyamila v. Canada (Public Safety and Emergency Preparedness), 2018 FCA 22, [2018] 3 FCR 674, at para. 46; (2) that once a question is certified the appeal is at large, such that the Federal Court of Appeal (and, if leave is granted, the Supreme Court) may address all aspects of the decision: Baker v. Canada (Minister of Citizenship and Immigration), [1999] 2 SCR 817, at para. 12; and (3) that deference may be appropriate on some or all aspects of the decision, even the certified question: Kanthasamy v. Canada (Citizenship and Immigration), 2015 SCC 61, [2015] 3 SCR 909, at para. 44.

However, the requirement of deference creates difficulties where a question has been certified, as Mactavish JA explained in Camayo:

Certified questions generally raise questions of law, including, as in this case, questions of statutory interpretation. However, the questions, as phrased by the Federal Court, require a yes or no answer. This invites correctness review by this Court. That said, as described above, this Court is required to engage in reasonableness review on questions of statutory interpretation. This creates the possibility that, in some cases, this Court may find the RPD’s interpretation of a statutory provision to be reasonable, yet this Court may say something entirely different in providing its own view of the matter in answering the certified question—something that the Supreme Court expressly tells us not to do (at para. 40)

Her solution was to require the reformulation of certified questions. Rather than posing a yes-no question, the Federal Court should state the issue in terms of reasonableness:

The potential misfit between reasonableness analysis and the definitive correct answer required by a certified question can, however, be avoided if the Federal Court were to formulate certified questions in a manner that asks whether a particular statutory interpretation or approach is reasonable. In this case, the second and third questions, as stated, call for a correctness response. I would therefore amend them to ask whether the particular statutory interpretation or approach suggested by the question is or is not reasonable (at para. 44).

Some commentators have queried whether the emphasis on institutional design in Vavilov calls this settled law into question, particularly the importance accorded by the Supreme Court to the legislative decision to create a statutory right of appeal (yours truly here at p. 25 and Jamie Liew here at pp. 412-414).

As Professor Liew notes, in Vavilov, the magic word “appeal” is a “default indicator for correctness review”; and the Supreme Court itself once remarked that the certified question procedure “would be incoherent if the standard of review were anything other than correctness” (Pushpanathan v. Canada (Minister of Citizenship and Immigration), [1998] 1 SCR 982, at para. 43).

Indeed, Mactavish JA remarked in Camayo that the “solution” of applying correctness review on the legal aspects of a certified question “would seem to gain greater credence now that the Supreme Court has held that statutory standards can have a bearing on the standard of review” (Camayo, at para. 41). The broader context is relevant too: as the Board renders tens of thousands of decisions annually, with individuals’ ability to remain in Canada often at stake, the certified question regime can enhance certainty by ensuring that the Board takes a uniform approach on legal questions.

Pouring too much wine into the ‘institutional design’ bottle might prove problematic (as noted here at p. 39) but the unique features of the immigration regime could allow for correctness review where questions have been certified without having unfortunate consequences in other areas of law.

The approach to the formulation of certified questions in Camayo prompts a question about a different mechanism, raised by Stratas JA in Mason as an alternative means of permitting the federal courts to authoritatively resolve questions of law:

To avoid the prospect of duelling administrative interpretations of paragraph 34(1)(e)—and all the uncertainty, inconsistent application and unfairness that might result—administrators tempted to reach a different interpretation may wish to follow another route. At any stage during proceedings, a “federal board, commission or other tribunal”, such as the Immigration Appeal Division, may “refer any question or issue of law…to the Federal Court for hearing and determination”: s. 18.3(1) of the Federal Courts Act, R.S.C. 1985, c. F-7. Needless to say, in such a reference, the Federal Courts would not have to defer to any administrative decision-making, could receive all necessary evidence and submissions, and could pronounce the correct state of the law. Here, we have a major, stand-alone issue of legislative interpretation that arises on the facts of these cases. It is purely legal, requiring an examination of text, context and purpose. Unlike some administrators such as the National Energy Board in Forest Ethics above, the Immigration and Refugee Board does not have any particular expertise that might contribute to the analysis of text, context and purpose. In the future, a member of the Immigration and Refugee Board could well conclude that it is entirely appropriate to refer this issue to the Federal Court for resolution once and for all (at paras. 77-78, emphasis added).

The reference procedure is not encumbered by pre-Vavilov case law in the same way as the certified question procedure. There was no hint, for example, in Reference re Broadcasting Regulatory Policy CRTC 2010-167 and Broadcasting Order CRTC 2010-168, 2012 SCC 68, [2012] 3 SCR 489, that the standard of review was relevant at all to answering the question referred (see especially at paras. 11-13). Thus it may indeed offer a cleaner slate for correctness review than the certified question procedure, as Stratas JA suggested in Mason.

That said, in light of Camayo, it is not clear that the reference procedure would indeed proceed without any deference: perhaps it would be appropriate there, too, to reformulate the question in terms of reasonableness. Indeed, there is some recent case-law applying the reasonableness standard in the context of a statutory right to a “reference” (albeit to a statutory court rather than to a superior court): R. v. Joslin, 2022 ONCJ 151, at para. 27; The Province of New Brunswick (Chief Firearms Officer) v. Springfield Sports Club Inc., 2022 NBQB 56, at para. 16. In the latter case, Morrison J commented:

The statutory review is called a “reference”.  The question becomes whether the review process is the equivalent of a statutory appeal thereby rebutting the presumption of reasonableness as the standard of review.  In my view, it is not.

Ultimately, the question is whether an administrative decision-maker’s choice to refer a matter to the federal courts rebuts, in and of itself, the presumption of reasonableness review. This is certainly a possibility, though in the immigration context — to my mind at least — the certified question procedure seems tailor-made to achieve correctness review on questions of law.