Issue #199: January 2026
Plastics case, good character and professional regulation, remedies.
Dear readers,
Some broader news in the Canadian administrative law community, as a reminder:
As I pointed out last week, the Honourable Guy Régimbald of the Federal Court, John J. Wilson of Gowling WLG, and I are proud to announce the upcoming publication of 4th edition of LexisNexis Canadian Administrative Law is now available for pre-order. Pre-order now!
Earlier in January, the Federal Court of Appeal rendered its decision in the Emergencies Act case. I analyzed the case here.
Review of Executive Legislation
Canada (Attorney General) v Responsible Plastic Use Coalition, 2026 FCA 17
When the Supreme Court rendered its opinion in Auer v Auer, 2024 SCC 36 observers were worried that the application of Vavilovian reasonableness review would lead to judicial micromanagement of executive policy choices. For my part, I wasn’t so convinced. Properly applied, reasonableness review can be sufficiently deferential, particularly where broad legislative authority enables executive policy-making (see also Canadian Coalition for Firearm Rights v Canada (Attorney General), 2025 FCA 82).
The Federal Court of Appeal’s decision in Responsible Plastics is a good example of how Auer review is—and, in my own anecdotal experience, turned out to be—sufficiently deferential when the enabling authority calls for it. The Court’s decision (per Rennie JA) effectively shows how and why Auer review is keyed to the most important feature of the law of judicial review under Vavilov: the language of the enabling statute. While the Federal Court’s decision was, in my view, justifiable, Rennie JA’s decision supplies good reasons to think otherwise. But importantly, the Federal Court and Rennie JA are disagreeing about the right thing—what the enabling statute properly permits.
The case centres on the Governor in Council (GIC)’s issuance of an order under the Canadian Environmental Protection Act [CEPA], listing “plastic manufactured items” as a toxic substance. The Federal Court held that the government’s assertion of regulatory authority over a broad category of “plastics” did not fit the statutory scheme, for two broad reasons: (1) the CEPA distinguishes between the singular and the plural, and so it would not make textual sense to say that “plastics” as a genus is captured by the legislation; (2) not all plastic items in this broad category have the potential to be pollutive. Rennie JA recapitulates the Federal Court’s holding in the following way: “…not all plastics enter the environment, [and so] not all plastics cause harm; consequently, [the legislation] required the specific identification of the particular plastics that enter the environment…” [14].
Rennie JA, however, takes issue with both lines of reasoning adopted in the Federal Court’s decision. He starts by noting that the CEPA is structured in a two-step process. Step 1 permits the GIC to list a substance as toxic. The GIC must only “be satisfied” that the substance “may enter the environment” or “may cause harm.” If the GIC is so satisfied, and under Step 2, it can then regulate that substance, but that regulation requires “exquisite particularization” [13]. As Rennie JA notes, the effect of a listing decision enables the Ministers to further consider how to control the substance through regulatory or non-regulatory responses. This two-step structure “underpins the reasonableness inquiry into the GIC’s answers to those questions” [92].
Rennie JA concludes that the Federal Court’s decision effectively collapsed this two-stage process into one stage, because it required that specific plastic items be listed at the initial stage [13-14]. That belied the statutory scheme, which requires particularization at the regulatory stage. The effect of the Federal Court’s decision was to narrow the GIC’s power at the listing stage.
This meant that the Federal Court improperly required that the GIC list specific plastic items. At the listing stage, the language used by Parliament is permissive and discretionary. It permits the GIC to operate on a precautionary basis (“may” and “. As Rennie JA points out, the terms deployed at this stage—“substance” and “harm” are broad, open-ended, and qualitative in nature [110]. Rennie JA was not convinced that the singular/plural issue was decisive in the context of the two-step scheme, and given the Interpretation Act’s statement that words in the singular include words in the plural [55-57].
Rennie JA then went on to consider whether the GIC improperly concluded, based on the available evidence, that all plastics “cause or might cause harm” [142]. He rejected this argument. It is not that all plastics cause harm, but that all “plastic manufactured items” have the potential to cause harm, and that regulatory authority may be required to specify that listing [146-149]. To the extent the Federal Court held otherwise, it engaged in a “line-by-line treasure hunt for error” [149].
I view this case as proof-positive of a few facts about the law of judicial review in Canada. First, Auer review assesses deference at the source—through the interpretation of statutory language. Indeed, the disagreement between the Federal Court and Rennie JA in this case is a disagreement about what the statute means. This is productive ground on which to disagree. Prior to Vavilov, the disagreement might have revolved around a more abstract and value-laden question: whether the executive should receive deference when it engages in a polycentric type of decision-making. That is an ancillary question to the main one, which is whether the legislation contemplates such deference. The Federal Court thought it did because of the precision of the word “substance.” Rennie JA, however, took a more contextual view of the statute, noting how its machinery operates. While the Federal Court’s decision is justifiable, Rennie JA offers good reasons to doubt it, especially because—as he notes—the language used in the statute (“satisfied” “may” “harm”) are qualitative in nature.
Second, Auer review is not insufficiently deferential. As I noted above, there were worries that Auer review would permit the court to wade into the record to identify alternative regulatory options that might better achieve some stated regulatory purpose. As this case shows, Auer review counsels precisely the opposite. Rennie JA, under Auer review, rejected the invitation by the respondent to “shift the debate away from the language of the statute by asking whether the evidence could have been stronger, whether laboratory tests should have been conducted or whether only those items that might ultimately be regulated should have been listed in the Order” [78].
As Rennie JA points out, those are the wrong questions. Instead, the question is “straightforward”: “whether…the GIC could have reasonably been satisfied…[that] PMI have the potential to cause harm to the environment” [78]. That is a legal question which asks whether there is some rational basis in the record to conclude that the GIC could be “satisfied” as to a relevant set of facts (here, the potential for harm). That is different than a court faulting the GIC for choosing one regulatory option over another, even though both of those options are plausibly supported by legislation.
Professional Regulation and Statutory Interpretation
Law Society v Ontario v AA, 2026 ONCA 47
Professional regulation is an area in which courts have generally taken a deferential approach, particularly when assessing penalties. That approach has generally survived Vavilov, as James Casey aptly pointed out in his contribution to the Vavilov at 5 symposium. But there are some interesting signs that the post-Vavilov jurisprudence could have interesting spin-off effects in the world of regulation. Specifically, the Supreme Court’s decision in Mason v Canada (Citizenship and Immigration), 2023 SCC 21 (Issue #108) .meaningfully increased the scope of what decision-makers must do to properly interpret statutes. Moreover, Mason stands for the proposition that courts will hesitate before declaring that a decision-maker offered an implied interpretation of a statutory provision.
All of these themes are present in the Court of Appeal for Ontario’s (per Sossin JA) recent decision in AA. This case surrounded a good character finding made by the LSO. In 2009, while living abroad, the applicant sexually abused three children, including his eldest daughter, through clothed sexual touching. He later disclosed his actions but was not criminally charged. Between 2009 and 2017, the applicant was dishonest with medical practitioners, child protection authorities, and the Law Society of Ontario about his misconduct. In 2012, he applied for a licence but withdrew after an investigation. He reapplied in 2019, prompting a hearing into whether he met the “good character” requirement.
The LSO Tribunal Hearing Division (upheld on appeal at the Appeal Division), found the applicant to be of good character and imposed a licensing condition requiring supervision in meetings with minors.
The Court of Appeal found this conclusion unreasonable, fundamentally because of a flawed statutory interpretation analysis [123]. It is useful to explore Sossin JA’s analysis, in part because it presents a defensible reprisal of the Supreme Court’s recent statutory interpretation jurisprudence.
The case revolved, principally, around s.27 of the Law Society Act. Section 27 (2) merely states that “[i]t is a requirement for the issuance of every licence under this Act that the applicant be of good character,” while s.27(4) mandates that a licence may be refused only after a hearing by the Hearing Division. Following the Supreme Court’s guidance in CISSS A, 2024 SCC 43, Sossin JA pointed out that this text is “the most important constraint on the Hearing Division’s interpretation of s.27” [52]. As I have written before, this means that a court engaging in statutory interpretation must, first, notice whether the text is capacious enough to warrant enrichment via context and purpose. Here, as Sossin JA pointed out, “…the text of s.27 is broad and open-ended” [51]. Given the lack of definition in the statutory terms (see, recently, the Supreme Court’s explanation in Lundin Mining, 2025 SCC 39 at para 6), this meant that context and purpose were dominant considerations in the case.
The Supreme Court’s approach to administrative interpretations of law in Mason calls for fairly clear engagement with the text, context, and purpose of the statute, and so the question here was whether the LSO Tribunal adequately grappled with context and purpose, which took on a greater role given the relative generality of the text. As Sossin JA points out in this case, “[a] key context for the good character assessment is that of the self-regulation of the legal professions in the public interest…” [61]. In other words, the Tribunal had to turn its mind to whether its good character finding was consistent with the broader public interest mandate of the law society [117-118]. This was a straightforward conclusion: given the importance of context and purpose here, its omission from the analysis was fundamental.
There is a lot of interesting detail in Sossin JA’s decision that is worth considering. First, Sossin JA’s treatment of the text in this case offers a good example of how CISSS A’s “text as anchor” approach works. In cases of textual precision, there will be less room for supplementation and enrichment via context and purpose (see, most recently, the Supreme Court’s decision in Kosicki v Toronto (City), 2025 SCC 28 at para 37). In other words, Sossin JA was right to give attention to the public interest context/purpose on this particular occasion.
But, secondly, there are risks with this approach. There are two associated worries. First, under the text as anchor approach, the Supreme Court has offered little guidance (so far) about when a text is sufficiently general to fault a decision-maker for failing to address context and purpose in a proper manner. This means that judges may disagree about when a text is sufficiently general. This was the precise disagreement in Reference re iGaming Ontario, 2025 ONCA 770 (Issue #198).
Answering this question is difficult, but it should be sufficient to say that consideration of context and purpose is relevant in every case. In other words, the question is not whether the text is “clear” enough to warrant addressing context and purpose—that would be a methodological backslide to an era where courts sometimes considered whether meaning was “plain” enough. Instead, the question is the uses to which context and purpose can be put in relation to text. For example, a precisely-drafted text will resist efforts to add on “omitted exceptions” to the semantically-bound statutory term (Kosicki furnishes a good example). One can contrast such a situation with AA, where the good character finding must be constrained by considerations of context and purpose.
Secondly, Sossin JA’s contextual analysis calls to mind the Supreme Court’s recent steeling of reasonableness review in cases like the Cabinet Mandate Letters Case, 2024 SCC 4 (Issue #123). Recall, there, that the Supreme Court majority faulted Ontario’s Information and Privacy Commissioner for failing to consider the constitutional significance of Cabinet privilege in its decision to compel the production of Cabinet mandate letters. Karakatsanis J took a bold step (an unjustified one, by my lights) to hold the Commissioner to a definition of Cabinet privilege that arose out of academic work [29-30]. There is a risk that, under the guise of robust reasonableness review (and an expansive approach to “context”), reasonableness review will become a tool that requires run-of-the-mill decision-makers to grapple with quasi-constitutional norms without warrant or prompting from a party or statutory context.
This worry is live on the current state of the Supreme Court’s jurisprudence, but AA is not an example of this “contextually-heavy” form of reasonableness review. In the AA case, the public interest is littered throughout the statutory context—it does not just arise from the quasi-constitutional ether or academic work. Reasonableness review must walk a fine line. It is one thing—as in Mason and AA—to require a decision-maker to reason through statutory context and textually-recorded statutory purposes. That is a function of the higher bar erected by Vavilov. It is quite another to force a decision-maker to address unargued Charter values, or academic treatises. In this sense, AA walks the line well.
Remedies
One of the more curious aspects of the post-Vavilov landscape: in every post-Vavilov case, the Supreme Court has found only one reasonable result on the law and directed the proper result. This approach to remedies, of course, is entirely inconsistent with the premise and function of reasonableness review. As Rowe J shrewdly noted in his concurring opinion in Pepa v Canada (Citizenship and Immigration), 2025 SCC 21 this approach takes reasonableness review far beyond where it should go.
Luckily, lower courts—as they sometimes do in Canadian administrative law—have limited the impact of the Supreme Court’s jurisprudence. Two recent Federal Court of Appeal cases are apposite.
In Canada (Attorney General) v 1230890 Ontario Ltd, 2026 FCA 4, Stratas JA declined an invitation to direct a particular result. The Attorney General—undoubtedly bolstered by the Supreme Court’s approach—sought this outcome. It is worth outlining Stratas JA’s reasons for rejecting the invitation:
[11] The Attorney General submits that, based on the evidence in this record, there is only one possible result in this case and so we should direct the Tribunal to confirm the notice of violation, including its imposition of penalty. We disagree. As the Supreme Court said, such a direction—in effect a mandamus order—is to happen only in “limited scenarios”: Canada (Minister of Citizenship and Immigration) v. Vavilov, 2019 SCC 65, [2019] 4 S.C.R. 653 at para. 142; and see the high threshold for mandamus discussed in, e.g., Canada (Public Safety and Emergency Preparedness) v. LeBon, 2013 FCA 55 at para. 14. The Tribunal is the merits-decider under this legislative regime and departures from that should be rare: Immeubles Port Louis Ltée v. Lafontaine (Village), 1991 CanLII 82 (SCC) , [1991] 1 S.C.R. 326 at 361 ; Maple Lodge Farms at para. 52; Terra Reproductions Inc. v. Canada (Attorney General), 2023 FCA 214 at para. 8. The Attorney General has not persuaded us that the evidence unequivocally points to violation. It may also be that, given the clarification of the legal matters in this decision, if the matter is remitted to the Tribunal for redetermination the respondent might make submissions on liability and penalty that were not open to it before or that were not properly understood or considered before, with the effect that the result and supporting reasons might change. Thus, in our view, the high threshold for directing a verdict in this case has not been met.
Similarly, in Westjet v TH, 2026 FCA 20, Gleason JA similarly rejected such an invitation:
[35] I turn finally to the appropriate remedy. While this Court possesses discretion to direct a determination on the merits of a complaint such as this where it finds that the Board’s decision is unreasonable, this discretion should not be lightly exercised. As noted by the Supreme Court of Canada in Vavilov, “… it will most often be appropriate to remit the matter to the decision maker to have it reconsider the decision, this time with the benefit of the court’s reasons” (at para. 141). Here, there is no forgone conclusion as to what the appropriate outcome is since dismissal is not necessarily warranted in every instance of sexual harassment: Harris at § 5:60; Brown & Beatty at § 7:33. It is therefore for a Board adjudicator and not for this Court to redetermine the respondent’s complaint.
Both cases are important because they erect a high-bar for a directed outcome. Consider the language used in these cases: the discretion should not be “lightly exercised”; it requires a “high threshold”, with departures from the regular rule of remittal being “rare.” One can contrast this language with the best (though by no means satisfactory) explanation the Supreme Court has offered for its own practice in Pepa. There, Martin J seemed to strike a different chord, noting that “[i]t should not be a surprise that there are cases in which one reasonable interpretation may arise, given that when legislatures speak, they intend to speak with clarity and purpose” [125]. Given—in fact—that there is a presumption of consistent expression (legislatures speak consistently because they speak with clarity and purpose), a court following Pepa’s lead might be much more comfortable simply directing outcomes.
In terms of fundamental principle, the Federal Court of Appeal’s approach provides a better account of what courts should actually do on judicial review. One gets the impression that the Supreme Court’s direction of “one reasonable outcome” in cases like Pepa is driven by its appreciation of the equities of the particular case—a felt need, perhaps, to avoid absurdity or injustice in relation to particular claimants. But fundamental doctrine should never be derived from results-oriented reasoning. The law of judicial review is structured in a particular way: statutory delegates are granted the power, at first instance, to render decisions on the merits. That power remains even when the delegate exercises the power unreasonably. While there are exceptions to this rule, the Federal Court of Appeal is right to keep those exceptions narrow in order to preserve the fundamental structure of administrative law. Judges on judicial review are not free-roving institutions designed to vindicate rights in the abstract. Their powers are nested within a structure that cabins and channels the judicial review function in order to preserve delegated power.
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.