Simplicity in the Law of Judicial Review of Regulations: Auer and TransAlta

This post is derived from this week’s edition of my newsletter, the Sunday Evening Administrative Review.

______________________________________________

Auer v Auer, 2022 ABCA 375 (November 22, 2022); TransAlta Generation Partnership v Alberta (Minister of Municipal Affairs), 2022 ABCA 381 (November 23, 2022)

Context and Holding: In these decisions, the ABCA deals with the question of how courts review regulations for compliance with primary law. The cases hold that the framework set out in Katz Group Canada Inc v Ontario (Health and Long-Term Care), 2013 SCC 64 applies, rather than the revised judicial review framework in Vavilov. In so doing, the ABCA sets itself up directly opposite from the Federal Court of Appeal, which has endorsed Vavilov as the starting point for the review of regulations: Portnov v Canada (Attorney General), 2021 FCA 171 (Issue #7). It is also set up opposite the BCSC/ BCCA: see e.g. Pacific Wild Alliance v British Columbia (Forests, Lands, Natural Resource Operations and Rural Development), 2022 BCSC 904 at paras 68-75; Whistler (Resort Municipality), 2020 BCCA 101.

Analysis: As readers of this newsletter will know, I strongly disagree with the ABCA’s conclusion, and the reasons underlying it. Reading Auer and TransAlta together, the ABCA advances several reasons for preferring the “hyper-deferential” Katz framework for the review of regulations over Vavilov:

1. Vavilovian reasonableness impermissibly invades the exercise of legislative powers, violating a core tenet of the separation of powers. Katz “maintains the integrity of the separation of powers and the role of the legislative branch of government. It ensures that courts do not enter the legislative field by weighing in on matters that properly fall within the sphere of the legislature and the executive” (Auer, at para 58; see also para 83). Or as put in TransAlta, at para 50: “[t]o decide whether a valid regulation is, in outcome ‘reasonable’ is to judge the merits of the path chosen by the delegated lawmaker to achieve the objectives of the enabling statute.”
2. Vavilov cannot be taken to implicitly overturn Katz: it makes only one passing reference to Katz (see Auer, at para 42; TransAlta, at para 47).
3. There are practical problems with applying Vavilovian reasonableness review to regulations. As Auer notes, “…[m]any of the contextual factors highlighted in Vavilov simply have no application to a vires review” (Auer, at para 77).

I will respond to these three concerns, but I first want to highlight a core feature of my response. There have long been debates in the law of judicial review over the need to “limit and simplify” versus the need to “tailor deference to variety” (US v Mead Corp, 533 U.S. 218 at 236). Of course, this is rarely a binary, and because of the subject matter, some consideration of variety will be necessary (as Vavilov‘s acceptance of context demonstrates). This is inevitable. Nevertheless, I am on the side of limitation and simplification to the extent possible. The fact that administrative decision-makers come in all shapes and sizes does not mean we require legal rules that track every individual type of decision-maker or decision, absent any fundamental reason. To my mind, all that is required is: (1) the recognition of fundamental principles that guide the doctrine (in Vavilov’s case, legislative intent and the rule of law); (2) the creation of general, all-purpose doctrinal rules plausibly connected to these principles; (3) guidance on how to apply the doctrine.

Vavilov and its progeny accomplish this. While Vavilov is, admittedly, contextual, it simplifies judicial review because it provides (1) a set of standard of review categories that plausibly map to legislative intent and the rule of law (though imperfectly); (2) on the reasonableness standard, it provides guidance about the contextual constraints that are relevant in a given case–this guidance limits these constraints so courts and litigants know when they will be relevant. Most importantly, when I speak of simplicity, I think of the fact that Vavilov provides an agreed-upon starting point, connected to fundamental principles, for all review of action of all kinds taken under delegated power. In this sense, Vavilov is a hard-won template. As we will see, the recent case of Law Society of Saskatchewan v Abrametz, 2022 SCC 29 (see Issue #48), inexplicably unmentioned by the ABCA, endorses the “start with Vavilov” idea on a question outside Vavilov’s contemplation: procedural fairness. This shows Vavilov’s utility as a general framework.

Starting with the same well of conceptual resources for all sorts of decisions simplifies the law of judicial review, and is no small thing. Simplification isn’t just aesthetic. Lawyers—to their detriment—sometimes overcomplicate matters beyond what is necessary, perhaps out of academic self-satisfaction. But the reality is this: the law of judicial review must be workable. It must connect to fundamental principles but at the same time be applicable by judges and understood by parties who bear the brunt of state action. This is the gargantuan challenge of administrative law. In this sense, Vavilov has done an extraordinary thing by largely accomplishing this goal. Parties now tend to argue about the merits of their cases rather than the standard of review. The ABCA’s discursus on Katz, unfortunately, is a step back to the old days of distinctions between legislative/quasi-legislative/adjudicative functions, where there are islands of government power uninhibited by the regular law of judicial review. If there was a compelling reason in principle for this, that is one thing. In this case, the Court’s decision endorses Katz because of its own erroneous perception of what the separation of powers, Vavilov, and general principles of administrative law require.

On to some specific points of contention:

1. The ABCA’s separation of powers argument does not get off the ground because of (1) a fundamental (though understandable) confusion about the word “merits” in Vavilov; and (2) a confusion about the role of secondary legislation. As Paul Daly argues, (1) leads the ABCA astray. Auer says that “[a] true Vavilov approach can only be accomplished by the reviewing court descending into a consideration of the merits of the policy decisions underlying the regulations and formulating its own reasons why the regulation was a reasonable policy choice” (Auer, at para 75). As Daly says, it is true that Vavilov speaks of its framework applying to the “merits” of administrative decisions (e.g. Vavilov at paras 2, 10, 16). But this does not mean that Vavilov endorses a judicial questioning of the policy wisdom of an administrative decision. This simply cannot be the case as a matter of fundamental principle. Vavilov’s reference to merits, instead, refers to the substance of administrative decisions as opposed to procedural concerns. As is well-known, judicial review polices the boundaries of the administrative state according to the concepts of legality, reasonableness, and fairness. This is different than questioning the policy merits of an administrative decision in the abstract. Judicial review—and Vavilov reasonableness—does not mean that courts arrogate to themselves the right to make certain policy choices. A few specific examples are relevant to show how this works throughout the law of judicial review:

• In Alberta Teachers’ Association, 2011 SCC 61, the Supreme Court addressed the situations in which it would be appropriate for litigants to make new arguments on judicial review. Generally, the presumptive rule is that new arguments cannot be made on judicial review, because “the legislature has entrusted the determination of the issue to the administrative tribunal” (Alberta Teachers, at para 24). This is a recognition that judicial review cannot proceed as a trial de novo, a recognition of the space left to the decision-maker to flesh out the law in its field so long as the decision fits within the purview of the statute.

• As the Federal Court of Appeal has stated with reference to new evidence on judicial review, the same rule applies because the legislature delegated the power to the administrator to “determine certain matters on the merits”; permitting new evidence routinely would undermine the demarcation between legislative and judicial roles, and so “[t]his Court can only review the overall legality of what the Board has done, not delve into or re-decide the merits of what the Board has done” (Association of Colleges, 2012 FCA 22 at paras 17-18).

Deference under Vavilov takes on a similar hue. Courts do not reweigh the evidence on judicial review (Vavilov, at para 125); deference necessarily involves a restriction on the court in intervening with an administrative decision-maker because that decision-maker has been delegated the power to make decisions (Vavilov, at para 13). These decisions may have policy consequences, but courts do not second-guess those consequences; they only ensure that a particular decision fits within the purview of the statute, and meets the basic requirements of rationality. This is even so where regulations are made after submissions in a legally-defined process: in such a case, the submissions form part of the record that courts use to assess whether the secondary legislation is justified by the law and the facts to which it applies. None of this involves, properly applied, an impermissible intrusion into the realm of lawmaking because the court is not formulating policy alternatives nor weighing in on which alternatives are best. It is only asking whether the action fits the bounds of the law and the evidence, like it does for all executive action. Indeed, this is the same rule we apply to all acts taken by the executive under statutory authority, including municipalities and other bylaw-creating bodies. As I will point out, the sweep of Auer/TransAlta is unknown, and presumably it should capture these bodies as well.

This is related to the second problem. The ABCA skirts over what I consider to be the real issue: the subordinate nature of what we call “executive legislation.” The ABCA’s entire point apparently seems to rest on the assertion that regulations are part of the primary legislative process—that, legally, the exercise of legislative powers by the Governor in Council is subject to the same rules that apply when Parliament enacts laws as an exercise of primary legislative authority (Auer, at para 53, citing Mikisew Cree First Nation v Canada, 2018 SCC 40 at para 32). Primary legislative authority is the authority to “enact, amend, and repeal statutes” (Pan-Canadian Securities Reference, 2018 SCC 48 at para 76). These statutes cannot be reviewed except for constitutionality, and this was the context of Mikisew Cree (notably not executive legislation). But this ignores a fundamental distinction between primary and secondary legislation. Inexplicably, the ABCA recognizes that regulation-making is “an act incidental to the legislative process” (Auer, at para 56), but does not take this to the logical conclusion. Secondary legislation (regulations) is subordinate legislation, which must fit the terms of the primary legislation. But regulations can be reviewed in order to determine whether they fit the scope of their enabling statute. Read literally, Auer seems to prove too much: if one simply transposes, as Auer does, the primacy of primary legislative authority to secondary legislation, one is endorsing a “hands-off” approach in judicial review altogether when it comes to executive legislation. But as we know, even if we follow Katz, secondary legislation can be reviewed to determine its fit with the governing statute, and so the analogy Auer draws to the primary legislative process is inapposite.

The point here is that secondary legislation is still executive action, amenable to review like all executive action–with the caveat that because of the legislative form of the action, it will be reviewed under Vavilov in a certain way (see point #3 below).

2. The “Vavilov does not mention Katz” argument has been made before: see e.g. Ecology Action Centre v Canada (Environment and Climate Change), 2021 FC 1367, and this post from Martin Olszynski and I. We were not impressed with this argument at the time, and I remain unimpressed, for two reasons. First, the “Vavilov does not mention x” argument has lost a lot of steam after Abrametz. Again, Abrametz held that questions of procedural fairness that arise under a statutory right of appeal are reviewed under the appellate standards. These questions were not mentioned in Vavilov. Following Abrametz leaves the ABCA on shaky territory. Second, as Prof. Olszynski and I wrote, the question is not whether this or that case was mentioned. Vavilov tells us where (as with Katz) there is a question as to the appropriate standard of review, a court “should look to these reasons first in order to determine how this general framework applies to that case” (Vavilov, at para 143). The ABCA in Auer and TransAlta do not even attempt to do this. This is despite the fact that Vavilov is a “holistic” framework (Vavilov, at para 143), one that is “sweeping and comprehensive” (Portnov, at para 25). Auer and TransAlta suggest that courts use a magnifying glass to see if particular examples of executive action are mentioned within Vavilov. This is unnecessary. They simply need to follow Vavilov’s general principles, as outlined in Vavilov, at paras 143-144.

3. The practical problems of applying Vavilov to regulations, with respect, do not exist. When Auer maintains that some of the legal and factual constraints listed in Vavilov do not apply in cases of regulation, the Court appears to misunderstand how Vavilov works. Not all of the constraints have to apply in a given case for Vavilov to be relevant. In some visa decisions, for example, statutory interpretation will not be the forefront consideration—these cases generally turn on evidence and findings of fact. With regulations, the dominant constraints will be the legal ones mentioned in Vavilov, and in many cases, deference will be expansive. This is not a surprise, though perhaps it is to the ABCA, which erroneously sees Vavilov as a more intrusive standard, always and everywhere (Auer, at para 61).

But the Court is also is too quick to discard the other constraints because it focuses on only one type of regulation-making: secondary legislation of general application. But as Portnov shows, this is not all there is. Portnov concerned the Governor in Council’s ability to “issue an order or regulation restricting or prohibiting any dealings with certain property held by designated individuals,” eighteen in total (Portnov, at paras 3,5). In such a case, the mere fact that the Governor in Council proceeded by secondary legislation does not immunize it from review on Vavilov grounds. In such a case, the statutory prequisites to the exercise of the power will be central. But because the court must also discern how the Governor in Council understood the authority granted to it under the primary statute (ie) to apply a regulation in these limited circumstances, the record must disclose the Governor in Council’s basis for its legal conclusion as applied to these individuals. In other regulatory cases, determining whether the regulation is justified by the primary law will depend on what explanations find their way into the record. This is the nature of Vavilov review, which is not always and everywhere more aggressive than Katz. Indeed, when we apply this review in cases of other “legislative” bodies including law societies and municipalities, the review looks fairly deferential, respecting the legislative posture of these bodies. While the Court calls this state of affairs “confusing,” (TransAlta, at para 49), I beg to differ: the same contextual constraints from Vavilov apply, with different force depending on the decision at issue. Regulations, if they are not primary legislation, are similarly nothing special as executive action.

I could say more—I hope to in longer form soon. But I end where I began. The ABCA’s approach will complicate the law of judicial review, not just because of its endorsement of a carveout for Governor in Council regulations. We do not know how far this could go. Are regulations made by agencies with a responsible Minister also captured by this rule? The logic should follow—and yet it would be a stretch to say that agency law-making is the same as primary law-making, especially given the deficiencies in the scrutiny of regulations process. What about rules of binding “legislative” effect created by agencies? These are unanswered questions left open by these decisions. The bottom line: when in doubt, start with Vavilov.

Paul Daly
John Mark Keyes