SEAR subscribers may be interested in reading a paper I wrote that will be published in the Canadian Bar Review soon. The paper is entitled “Foxes, Henhouses, and the Constitutional Guarantee of Judicial Review: Re-Evaluating Crevier.” The paper deals with a problem I’ve surveyed a few times in this newsletter: the scope of the constitutional guarantee of judicial review in the post-Vavilov world (see Issue #4). I hope the paper does justice to a very tough problem.
Mason v Canada (Citizenship and Immigration), 2023 SCC 21 (September 27, 2023)
Category: Selection of standard of review//application of reasonableness standard.
Context: I addressed the Federal Court of Appeal’s decision in Issue #3. Section 34 (1) (e) of the Immigration and Refugee Protection Act provides that permanent residents or foreign nationals are “inadmissible on security grounds” for “engaging in acts of violence that would or might endanger the lives or safety of persons in Canada.” The key question is whether the acts of violence had to be connected to national security, or whether violence that affects individuals can qualify under the provision. The Immigration Appeal Division held that the provision was broad, reaching beyond cases implicating national security. Mason et al sought judicial review in the Federal Court, which quashed the decisions. On appeal, the Federal Court of Appeal (per Stratas JA) decided that the IAD’s decisions were reasonable.
Issues: (1) What is the standard of review? (2) Are the decisions reasonable?
Holding: Jamal J wrote for a majority, allowing the appeal.
(1) The standard of review is reasonableness: “No established exception to the presumption of reasonableness review applies, nor should any new exception be created on the basis that the appeals involved a serious question of general importance certified for appeal to the Federal Court of Appeal under s.74(d)” [9].
(2) The IAD decisions were unreasonable. Principally, the IAD failed to consider issues raised in Mason et al’s submissions pertaining to the legislative context. The IAD also failed to interpret and apply s.34(1)(e) “in compliance with international human rights instruments to which Canada is a signatory…” [10].
Côté J concurred but differed on several points. She would have recognized a new correctness category: when appellate courts decide a ‘serious question of general importance’ under the certified questions regime in the Federal Courts [127]. She agreed with Jamal J and the majority that, on the merits, s.34(1)(e) “requires a nexus between the relevant act of violence and with national security or the security of Canada” [124].
Analysis: To make a long story short, I believe that this case makes two points: (1) it will take a special case to warrant the recognition of a new correctness category; and (2) the reasonableness standard will insist on a high bar for decision-makers, within the bounds of appropriate deference, especially when determining whether a decision-maker dealt with (a) statutory constraints and (b) submissions raised by the parties. There are also some unfortunate missteps by the Court, in my view, that lead to odd contradictions in the decision, contradictions that I hope will be worked out, but which are concerning nonetheless.
First, the debate in this case over the recognition of a new correctness category raises the difficult question of when to contemplate “edits” to the Vavilov framework. On one hand, Jamal J leans heavily on Vavilov’s commitment to “simplicity, coherence, and predictability to the law on standard of review” [44]. For those who grew up in the Dunsmuir era, Jamal J’s next line will come as happy news: “…reviewing courts need no longer conduct an unwieldy ‘contextual’ inquiry to identify the appropriate standard of review, as had been the case in the period before Vavilov” [44, see also 53]. In part, this motivates the conclusion that the certified questions regime cannot be analogous to other correctness categories. Here, “[c]ertification of a question under s.74(d) of the IRPA does not signal that the legislature intended that the appellate courts apply correctness review,” especially because in Vavilov itself, a certified question was involved under the Citizenship Act [52].
The goal of simplicity and coherence in the law of judicial review is an important takeaway—perhaps the important takeaway—from Vavilov. It is a theme I have struck on many occasions, including on cases involving the standard of review of subordinate regulations. It is why the Court should be hesitant to recognize new correctness categories, or similarly, introduce new “general questions of central importance.” Most courts in the post-Vavilov world have taken this warning seriously.
Nonetheless, I am attracted to Côté J’s reasoning on the certified questions category, in part because I think the recognition of a correctness category here is narrow and consistent with Vavilov. It seems to me that this case is quite analogous to SOCAN v ESA, 2022 SCC 30 (Issue #49). There, there was a legislative signal “as strong and compelling” as a legislated standard of review or statutory right of appeal: a scheme of concurrent jurisdiction (Vavilov, at para 70). The certified questions regime seems to do the same thing, signalling a “legislative intent for judicial involvement” (SOCAN v ESA, at para 31). Why is this so? As MacTavish JA cogently argued in Galindo Camayo, 2022 FCA 50, the certified questions regime contemplates a court providing yes/no answers to specific questions; it invites judicial involvement in the IRPA by directing courts to provide clear answers to certain, defined questions. This seems on the level with SOCAN v ESA, or rights of appeal, in that it is a legislative signal for correctness review and the equal involvement of the courts in the statute. Jamal J seems to be searching for a level of explicit expression in legislative signals that Vavilov did not require when he does not see this aspect of the certified questions regime as a compelling signal.
Jamal J may have been motivated by other considerations, but none of these limit the force—in my view—of Côté J’s basic point. For one, it is true that the certified questions regime is also a “gatekeeping” mechanism, which might suggest that it only conditions the circumstances in which appeals can be sought rather than specifying a standard of review [49]. Vavilov suggested that some statutory terms that appear to offer appeals will only condition the extent to which judicial review can be sought (Vavilov, at para 51). But it is hard to gainsay what MacTavish JA mentioned in Galindo Camayo: the certified questions regime does change the substantive function of courts answering those questions, since the regime contemplates the definitive judicial resolution of important legal questions.
Perhaps Jamal J was motivated by the ill-fit of the Federal Court of Appeal applying a correctness standard on a case containing a certified question, while the Federal Court would apply a reasonableness standard. But this is no more awkward than the current setup, which forces the Federal Court of Appeal to reformulate certified questions to suit the reasonableness standard. This seems to undermine the certified questions regime itself, which contemplates yes/no answers. Absent a reformulation, we’re left with a hypothetical situation where a certified question is answered in a certain way, but a tribunal’s decision—which may be different—is nonetheless upheld as reasonable.
Nonetheless, Jamal J holds the majority here, and so it is worth asking about implications of his approach. It is important to recognize the value of simplicity in the law of judicial review, but simplicity should also be consistent with principle. It is difficult to recognize the distinction between SOCAN v ESA and Mason. Perhaps this isn’t a big issue—there are only so many legislative arrangements inviting judicial involvement. And the bottom line of Jamal J’s opinion appears to be a reticence to recognize new correctness categories absent a compelling reason. To preserve the Vavilov compromise, this may require some of us who think Côté J had the best of Mason to put water in our wine.
Second, as I have outlined in this newsletter many times, courts across the country have generally interpreted Vavilov to provide more exacting reasonableness review on questions of law. The reasons offered by the decision-maker in such cases is the starting point of the analysis, and those reasons must “be consistent with the text, context, and purpose” of the provision under interpretation [120]. What this has amounted to is a requirement that decision-makers address contested aspects of the statutory context (see e.g. Whissell v Tribunal administratif du travail du Quebec, 2022 QCCS 1113, finding fault with the tribunal because it failed to consider statutory purpose; Canada (Attorney General v Hull), 2022 FCA 82, finding fault with the Social Security Appeal Division for rendering an interpretation that runs afoul of the text and purpose). The fine line courts must walk in conducting this more robust review is complicated because deference cannot disappear. Appeal courts in the post-Vavilov world have found fault with first instance judicial review courts that conduct “disguised correctness” review, a significant problem in the pre-Vavilov world (see e.g. New Brunswick Nurses Association, 2023 NBCA 60). And reasons need not be perfect, nor necessarily explicit; sometimes, we can read the record and determine that a decision-maker’s interpretation is justified (see e.g. Le-Vel Brands LLC v Canada (Attorney General), 2023 FCA 177).
It might be said that this method is incoherent, but for my part, I think it is workable, if imperfect. The problem becomes how robust the reasonableness review must be. With some notable exceptions, this case seems to err on the side of more robust review:
The “Lay of the Land”: let’s start with a notable exception, and one of the contradictions in Jamal J’s reasons. In conducting reasonableness review, the Federal Court of Appeal instructed that courts should start with the reasons, but should also take a preliminary “lay of the land” in order to understand the statute. For the FCA, this lay of the land was not designed to create a yardstick and measure the tribunal’s reasons against it. That would be disguised correctness review. Instead, I read the Court’s comment in Mason FCA as a common sense invocation of the normal rule of statutory interpretation: we must understand statutory terms in relationship to other sections of the statute. It’s what L’Heureux-Dubé J said so clearly in 2747-3174 Quebec Inc v Quebec, [1996] 3 SCR 919: “It is thus necessary to read and reread the entire statute in full before deciding how the term in question should be defined.” For this reason, I find it hard to understand why Jamal J would find fault with the Federal Court of Appeal’s adoption of this common sense point. This is especially so when Jamal J worries that such an approach might lead to “disguised correctness review.” The Court of Appeal expressly noted that this was not the purpose of the lay of the land approach, and the Court of Appeal is well-aware of the problem of disguised correctness review, having called the Supreme Court to task for this problem in the pre-Vavilov world!
Nonetheless, no good judge could or should properly do what the Supreme Court tells them to do, and herein lies the contradiction. Understanding statutory terms requires reading the statute in full, and it would be odd to take the Supreme Court literally in its instruction not to do this. This instruction runs afoul of ordinary rules of interpretation, which Vavilov tells us apply with just as much force in the administrative context. I expect and hope courts will still read the statute in full when understanding the context the decision-maker made a decision. Otherwise, I think there is a fair worry that courts may blinker themselves to key aspects of the context that can constrain or expand administrative power, distorting reasonableness review. And as I note below, Jamal J can’t help himself—of course he looks to the statutory context to determine the relevant constraints bearing on section 34(1)(e), as addressed by the parties in their submissions.
My advice: courts should do what Jamal J does, not what he says.
The Legal Errors: Looking to the “lay of the land” himself, Jamal J finds fault with the IAD’s conclusion that a nexus is not required to national security for three reasons: (1) the IAD, even though it applied “several recognized techniques of statutory interpretation”, failed to address elements of the statutory context; (2) the IAD failed to address the consequences of its decision; and (3) the IAD failed to address constraints imposed by international law, as incorporated in the IRPA [85].
Statutory Context: Jamal J finds that the IAD failed to address key elements of the statutory context of the IRPA (seemingly finding the lay of the land quite dispositive). Both elements involved other sections of the IRPA that arguably supported the conclusion that Parliament carved out a special standard of severity in s.34(3) for national security offences, as opposed to offences that relate to public safety generally. For one, relief on humanitarian and compassionate grounds is not available for those who fall under s.34(3), as opposed to those convicted of serious criminality, for example [88]. And the legal standards are different: when someone is found inadmissible for serious criminality, the Minister asks whether the individual poses a danger to the public in Canada. Inadmissibility for security asks the Minister to consider whether the individual poses a danger to the security of Canada. This suggests that the a nexus with national security may be required [92]. These points of context were raised by Mason et al and unaddressed by the IAD explicitly.
The Federal Court of Appeal concluded that the IAD impliedly considered these statutory constraints—this was an option open to it under Vavilov. But here, Jamal J quite evidently confirms that there is a high bar for determining that a decision-maker impliedly considered an issue (see para 97). This is consonant with the high bar constructed by some courts in the post-Vavilov world: see Zeifmans LLP v Canada, 2022 FCA 160 (requiring the court to be “in no doubt” that the Minister was aware of the interpretive concern); Cavendish Farms Corporation v Lethbridge (City), 2022 ABCA 312. For my part, it wasn’t obvious Vavilov would take this character, but I am glad it did. Courts should not try to draw blood from a stone and should simply find decisions unreasonable if a line of analysis cannot be found.
International Law: Jamal J faults the IAD for failing to address certain legal constraints imposed by international law, an issue the Federal Court of Appeal did not address because the argument had not been raised before the IAD [116]. The IRPA does quite expressly contemplate the Refugee Convention being a major part of the interpretive context [117]. Perhaps, as Paul Daly opines, the IRPA’s commitment to international law is a “uniquely powerful, textually explicit commitment to implementing Canada’s international law obligations in the IRPA, which is unlikely to have similar force in any other context (save, perhaps, citizenship).” This seems right. But in the meantime, we have an oddity. Jamal J faulted the Federal Court of Appeal for essentially falling into disguised correctness review. But he then decided an issue de novo, even though it is a question of law that would—normally—fall under the rules of reasonableness review. In deciding that question of law—if this is what the Court is telling us that the IRPA requires—judicial review courts will lack the appreciation and insights of the very decision-makers that Jamal J extols as experts, owed deference on questions of law [70].
Hence another contradiction. As Côté J says at para 172: “My colleague views the IAD’s “fail[ure] to address the legal constraints imposed by international law” as unreasonable (para. 104). With respect, I would have concerns, given the emphasis in Vavilov on a “reasons first” approach, with finding a decision to be “unreasonable” based on arguments that were not put before the administrative decision maker and that do not apply to the individuals actually before that decision maker.” Ideally, if we are committed to reasonableness review, we are in a pickle here: we must do what the Supreme Court says, but not what it does.
Mason is, for this reason, a mystery wrapped in a riddle. We achieve robust reasonableness review, yes, but the contradictions here go to the heart of reasonableness review itself. One hopes these contradictions don’t unravel the progress we’ve made in this area of the law.
**POST EDITED ON THE WEBSITE FOR FORMATTING**
Toth v Canada (Health and Addictions), 2023 FC 1283 (September 25, 2023)
Category: Judicial review of constitutional determinations
Context: This is a judicial review of several consolidated decisions made by the Minister of Mental Health and Addictions. The Minister refused requests for exemptions under s.56(1) of the Controlled Drugs and Substances Act. The relevant requests were filed by healthcare practitioners (HCPs) so that they “possess and consume raw psilocybin mushrooms in the course of their own professional training for psilocybin-assisted psychotherapy” [4]. In other words, the HCPs want to use psilocybin themselves as part of their training: “..for optimal results, qualified practitioners should have experienced with the psychedelic medicines that will be used to treat their patients” [6]. The Minister rejected the exemption requests, determining that the exemption was not necessary for “medical or scientific purpose[s] or otherwise in the public interest…” [14]. The Minister believed that the HCPs had an alternative path: to go through a clinical trial [14]. The applicants argued that the decisions could not stand for several reasons; particularly, for our purposes, they suggested that the decisions did not “address arguments about the impact a refusal would have on the HCPs’ and patients’ rights under s.7 of the [Charter]” [18].
Issue: What is the standard of review? Do the decisions satisfy the relevant standard of review?
Holding: There are several issues in this case. The Court concludes (1) that the standard of review for all issues is reasonableness and that (2) the decisions are reasonable.
Analysis: For those following the twists and turns of Doré review on constitutional issues, this is yet another entry in the anthology. Readers will remember that there are two lines of authority developing. The first line holds that a question of whether a decision-maker must consider the Charter in the circumstances of a particular case is a matter of correctness review, with the balance of Charter values with statutory objectives falling to reasonableness review (CBC v Ferrier, 2019 ONCA 1025). The second line of authority posits that a decision cannot be reasonable if a decision-maker fails to consider Charter arguments that are squarely raised in front of it (see Canada (Attorney General) v Robinson, 2022 FCA 59 (Issue #37). There are differences between these approaches, but the end result of both of them is the same. A decision that fails to consider Charter arguments that it should cannot survive judicial review,
The devil remains in the details. Here, the Court decides that “in applying the Dore framework, the Court should not carve out a freestanding question that an administrative decision maker must address, and which is subject to a correctness standard of review” [95]. This rules out Ferrier as a path in this case, and at any rate, that option was discounted by the Federal Court of Appeal in Robinson.
But the Court goes on to address the argument the applicants made relying on Robinson. Helpfully, the Court distinguishes pre-Vavilov cases that arguably allowed courts to find implied reasons where none existed [108]. Nonetheless, here, the Court concludes that “…the Minister’s failure to specifically mention the Charter does not render her decisions unreasonable based on the principles set out in Vavilov” [109]. This was because, according to the Court, “The Minister’s reasons provide the rationale for rejecting the Charter arguments by addressing their very foundation, which is that HCPs need experiential training to provide the most safe and effective care to patients” [112]. The HCPs argued that their liberty interests were engaged by the prohibition on possession, and that this prohibition not only deprived them of the ability to provide safe and effective care to patients, but also created a risk of imprisonment [112]. But as the Court says, the Minister rejected the premise that HCPs need experiential training in the first place to provide this care [112]. Having done so, according to the Court, the Minister actually did address the Charter arguments.
Two points on this. First, this approach will require a Court looking deeply into the reasons and the record to discern a line of reasoning that demonstrates that the administrator properly engaged the Charter issue. Maybe here this option was open to the Court. But I fear that this will only risk the prospect of courts diving into the record to discern something—anything—that can constitute “engagement” with Charter arguments. We recently saw this issue arise in the Peterson case (Issue #103). This seems far away from the joint rule established in Ferrier and Robinson, even if Robinson could be distinguished on the facts [113]. Second, the increasingly narrow and clever distinctions developing in this area of law underlines the need for major overhaul of the way we review exercises of administrative discretion engaging constitutional rights. While here the Charter was not engaged—this conclusion seems justifiable on the current state of the law—the “dancing on the heads of pins” that has characterized the law in this area as of late arguably demands another look by the Supreme Court.
Kumar v Canada (Citizenship and Immigration), 2023 FC 1279 (September 22, 2023)
Category: Application of the reasonableness standard.
Context: This is a decision of the Refugee Appeal Division which concluded that a family could safely seek refuge in India and therefore do not meet the definition of a Convention refugee under s.96 or a “person” requiring protection under s.97 of the IRPA. Among other things, the applicant family argued that the RAD failed to refer to another RAD case involving a refugee claimant fleeing police in Punjab [15]. The case at issue was released between the date the applicant family filed their written submissions and the date of the RAD’s decision, so the RAD was not made aware of the decision by the applicant family [16]. Nonetheless, the applicant family suggested that the RAD should have followed this case, or at the very least, addressed the other decision [16].
Issue: Is the decision reasonable?
Holding: Yes.
Analysis: I’ve highlighted this case because of McHaffie J’s remarks about the RAD. In rejecting the family’s argument, the Court says that the RAD is composed of “over a hundred members, who decide thousands of cases a year” [20]. In light of this, “it is effectively impossible for each member of the RAD to be aware of every decision issued by every other member” [20]. So says the Court, “[i]t is unrealistic to impose on the RAD, as the applicants propose, an obligation to be aware of all of its other decisions, let alone cite them or distinguish them, particularly in a case where they have not been raised” [20]. At any rate, in this case, the RAD case the family asserts should have been considered was not really “established” internal authority [21].
There may have been good reasons in this case to discount the RAD “precedent”; perhaps it was just not applicable in this case. And it is true that Vavilov does contemplate a reviewing court taking into account the operational realities of a given statutory regime. But I would caution about taking this too far. The tension between understanding administrative pressures and the requirement of justification can be difficult. Given the pressures all administrative decision-makers in Canada face these days, it would be unfortunate to limit the force of the requirement of justification in light of these pressures. That is not to say that this happened in this case. But ideally, we should not excuse failures of justification with reference to operational pressures.
Disclaimer: Nothing in the SEAR is legal advice. The SEAR is not designed to be comprehensive. It is designed to collect some cases that I believe are interesting to practitioners and students, and to provide my quick thoughts. Not all of the cases presented are fully analyzed ie) I’ve selected the most relevant issues in some of the cases. Please complement the SEAR with your own research and study. Any mistakes are my own.