Making a Monster

A report on the future regulation of the internet proposes giving the CRTC overwhelming and unaccountable powers

The final report of the Broadcasting and Telecommunications Legislative Review Panel, grandly entitled Canada’s Communications Future: Time to Act (the “BTLR Report”) has already attracted its share of commentary, much of it, but by no means all, sharply critical. As Michael Geist has explained, the report articulates

a vision of a highly regulated Internet in which an expanded CRTC … would aggressively assert its jurisdictional power over Internet sites and services worldwide with the power to levy massive penalties for failure to comply with its regulatory edicts. 

The discussion has mostly focused on the wisdom of the BTLR Report’s 97 recommendations for regulating the way in which Canadians engage with the online world, and also on their impact on freedom of expression. But one aspect of the report ― indeed, not merely an aspect but a fundamental element of the report’s underlying philosophy ― has, I think, received less attention, although Professor Geist alludes to it with his reference to “an expanded CRTC”: the report’s commitment to administrative power. This is, perhaps, a less obvious issue, but we should not underestimate its significance. If followed, the report’s recommendations would not merely expand the CRTC, but make into a bureaucratic behemoth. We must not let this happen.


The BTLR Report recommends multiple amendments to the legislation governing electronic communications in Canada that would tend to produce the “highly regulated internet” to which Professor Geist refers. Yet the striking thing is that most of the proposed changes do not describe the regulations that they call for with any precision. Instead, they say that the CRTC should be given vast powers to bring into being the report’s imagined brave new world.

The CRTC would be givens new powers to make rules of general application. Most ominously, it would be given the ability to regulate “media content undertakings” ― that is, all manner of entities creating their own content, whether written, sound-based, or visual, as well as those providing platforms for the content created by others, everything from a humble podcast to giants like Netflix, Facebook, and YouTube. These “undertakings” would be required to register with the CRTC, which would be

enable[d] … to establish classes of registrants, to amend registrations, and impose requirements — whether through conditions of registration or through regulations — on registrants (Recommendation 57)

These requirements could, in particular, include “codes of conduct, including provisions with respect to resolution mechanisms, transparency, privacy, and accessibility”. (Recommendation 74) At the same time, the CRTC would be given

the power to exempt any media content undertaking or classes of media content undertakings from registration in instances in which — by virtue of its specialized content or format, revenues, or otherwise — regulation is neither necessary nor appropriate to achieve media content policy objectives. (Recommendation 58)

In other words, the CRTC would decide ― with virtually no guidance from legislation ― both what the rules for “media content undertakings” would be an who would in fact have to comply with them at all. In particular it would be to

impose discoverability obligations on all audio or audiovisual entertainment media content undertakings, as it deems appropriate, including …  prominence obligations [and] the obligation to offer Canadian media content choices(Recommendation 62). 

The CRTC could impose similar requirements on “on media aggregation and media sharing undertakings” ― again “as appropriate” (Recommendation 73). The CRTC would also be directed to “intervene, if necessary … in order to respond quickly to changes in the communications services, improve transparency, and promote trust” in the face of technologies that “combine algorithms and artificial intelligence with Big Data” (Recommendation 93).

The CRTC would also be empowered, and indeed required, to regulate behaviour of individual market actors. It would be given the remit “to ensure that rates are just and reasonable” in “key electronic communications markets” (Recommendation 29). Indeed, in a rare instance of seeking to restrain rather than expand the CRTC’s discretion, the BTLR Report suggests that the ability of the CRTC to “forbear” from regulating the justness of rates should be eliminated (Recommendation 30). The CRTC would also be given the power to “regulate economic relationships between media content undertakings and content producers, including terms of trade” (Recommendation 61). In relation to CBC/Radio-Canada, the CRTC would be tasked with “overseeing all its content-related activities” (Recommendation 83).

But the report would not only have the CRTC make the law for the online world. It would also be given a substantial autonomous power of the purse. It would be given the power to designate “from an expanded range of market participants — all providers of electronic communications services — … required contributors to funds to ensure access to advanced telecommunications”. (Recommendation 25) Among the requirements the CRTC would be able to impose on those required to register … would be “the payment of registration fees” (Recommendation 57). It could, further, “impose spending requirements or levies on all media content undertakings, except those” mainly providing written news (Recommendation 61), “some or all” of which it could use to fund “to the production of news content” through “an independent, arm’s length CRTC-approved fund for the production of news, including local news on all platforms” (Recommendation 71).

The CRTC would acquire additional adjudicative powers too. For example, Recommendation 38 suggests that it should resolve disputes over the location of telecommunication infrastructure. More significantly, it would be both prosecutor and judge when “imposing penalties for any failure to comply with the terms and conditions of registration” imposed on “media content undertakings” (Recommendation 57), with “resolv[ing] disputes” among which it would also be tasked (Recommendation 61). Not that this adjudication would necessarily look like that done in the courts, since the BTLR Report would empower the CRTC “to issue ex parte decisions where the circumstances of the case justify it”. (Recommendation 75)

The prophet of the administrative state in Canada, John Willis, described administrative agencies as “governments in miniature”. One hesitates to describe the law-making, trade-regulating, money-grabbing CRTC envisioned by the BTLR Report as in any sense miniature, but it sure looks like a government unto itself, albeit a rather undemocratic one. In addition to the Commissioners who would exercise legislative, executive, and judicial powers, it would have a sort of representative body, the Public Interest Committee, “composed of not more than 25 individuals with a wide range of backgrounds, skills, and experience representing the diversity of public, civic, consumer, and small business interests, and including Indigenous Peoples”. (Recommendation 15) It’s not quite clear who would be appointing these people, but it certainly does not seem that, despite their supposed mandate to represent the public, they would be elected. Not to worry though: there would also be funding, out of fees collected by the CRTC, for “public interest interventions” (Recommendations 12 and 13), in case, I suppose, the Public Interest Committee doesn’t sufficiently intervene to represent the public interest. And, in addition to the prosecutorial and judicial functions of the Commissioners, there would be

an independent, industry-funded, communications consumer complaints office with the authority to investigate and resolve complaints from individual and small business retail customers of services covered by the respective Acts,

whose “mandate and structure” the CRTC would “create and approve” (Recommendation 96).

Meanwhile, outside control over this machinery will be be reduced. The Commissioners, who are currently appointed to renewable five-year terms, would instead serve for seven years, with no possibility of renewal (Recommendation 4). A limited form of Parliamentary supervision, the laying of government “directions” to the CRTC before the Houses of Parliament would be abolished in the interests of swift regulation (Recommendation 6). And, of course, given the vagueness of the legislative guidance to the CRTC and the breadth of its mandate, it is unlikely that the courts would intervene much to police its regulatory activities.

To sum up, the CRTC would be put in control, with very few restraints, of Canadians’ interaction with the online world, and with one another. Who can speak online and on what conditions ― the CRTC would have control over that. How much they have to pay for the privilege, and where the money goes ― the CRTC would have control over that. How disputes among them, and between them and the CRTC itself, are to be resolved ― the CRTC would have control over that too. The only “checks” on it would come from handpicked representatives of the “public interest” as the CRTC itself conceives it ― not from Parliament or the courts.


The empowerment of the CRTC proposed by the BTLR Report is, of course, no accident. It proceeds from a specific philosophy of government, which the Report describes quite forthrightly. According to its authors,

The role of government is to establish broad policies. The role of regulators is to implement those policies through specific rules and in a transparent and predictable fashion. Legislation is the key instrument through which government establishes these policies. It should provide sufficient guidance to assist the CRTC in the discharge of its duties, but sufficient flexibility for it to operate independently in deciding how to implement sector policy. To achieve this, legislative statements of policy should set out broadly framed objectives and should not be overly prescriptive. (46-47)

In other words, government ― Parliament is left out of the equation entirely, as if it has nothing to do with legislation ― should mostly leave the CRTC alone. Indeed, it is important to preserve “proper balance between the government’s role in policymaking and the regulator’s role in implementing those policies independent of government influence”. (47) And, judging by the amount discretion ― to make law and dictate the behaviour of individual organizations, to levy fees and spend money, to identify, prosecute, and condemn alleged offenders and to adjudicate disputes ― the BTLR Report would vest in the CRTC, the “balance” is really all on the side of the regulator.

This is the philosophy the BTLR Report would impose on the 2020s and, perhaps, beyond. It ostensibly envisions “the CRTC’s shift toward a future-oriented, proactive, and data-driven style of regulation”. (44) But its ideology comes, not from the future, but from a distant and, as article on “The Depravity of the 1930s and the Modern Administrative State” by Steven G. Calabresi and Gary Lawson about which I blogged here shows, detestable past. As Professors Calabresi and Lawson explain, President Franklin D. Roosevelt’s

administration and a compliant Congress created a vast array of new “expert” regulatory agencies, many of which followed the “independent” model by insulating the agency heads from at-will presidential removal, and many of which contained (and still contain) statutory authorizations to the agencies so vague as to be literally meaningless. … These agencies, controlled neither by the President nor by Congress, made life-altering decisions of both fact and law subject only to deferential judicial
review. (829)

This is the governance model proposed by the BTLR Report. Its original backers

fundamentally did not believe that all men are created equal and
should democratically govern themselves through representative institutions. They believed instead that there were “experts”—the modern descendants of Platonic philosopher kings, distinguished by their academic pedigrees rather than the metals in their souls—who should administer the administrative state as freely as possible from control by representative political institutions. (829)

(For more on the beliefs of 1930s pro-administrativists, see also this post by co-blogger Mark Mancini.) Judging by their proposals, the views of the authors of the BTLR Report are rooted in just this kind of thinking. They mistrust the free market as well as democratic institutions, and want fundamental decisions about what is, by their own account, an unbelievably important part of our lives to be made by officials deemed wiser than everyone else.

And if the philosophy behind the BTLR Report’s proposed future goes back a mere century, its institutional vision is considerably older still. In fact, at the risk of sounding a bit like Philip Hamburger (which, after all, isn’t a bad thing!) I would argue that it amounts to a counter-revolution against the 17th-century subjection of executive authority to law, and a reversal of the the post-1689 constitutional settlement. To be sure, everything the BTLR Report proposes to do would be covered by the fig leaf of ― deliberately vague and unconstraining ― legislative authority. But in substance, the proposals amount to executive law-making contrary to the Case of Proclamations, executive dispension from the law contrary to article 2 of the Bill of Rights 1688, executive adjudication contrary to the case of Prohibitions del Roy, and executive taxation contrary, this time, to article 4 of the Bill of Rights. James I and James II would be proud.


So when we hear that “this time it’s different” ― that the online world is like nothing we’ve seen before ― that its actors “pose a unique set of challenges for contemporary regulators”, as Paul Daly argues ― and that this justifies the sort of overwhelming regulatory response recommended by the BTLR Report, we need to be skeptical. For all that the issues raised by the modern world are ― now as a century ago! ― said to be quite unlike anything that came before, the solutions offered are the same old. More unfettered bureaucratic power is always said to do the trick. When all you have is a hammer…

More recently, a very different philosophy seemed, however briefly, to prevail in the online world. In the 1996 “Declaration of the Independence of Cyberspace“, John Perry Barlow proclaimed:

Governments of the Industrial World, you weary giants of flesh and steel, I come from Cyberspace, the new home of Mind. On behalf of the future, I ask you of the past to leave us alone. You are not welcome among us. You have no sovereignty where we gather.

The Declaration isn’t much more remembered than the term “cyberspace” itself, nowadays, and the weary giants whom Barlow was taunting have come after the cyber-libertarians like Pushkin’s Stone Guest. If the authors of the BTLR Report get their way, the we would indeed be governed, to keep with the 17th century English political thought, by Leviathan himself.


NOTE: A petition to “the Government of Canada to Reject the recommendations regarding the legislation and regulation of free speech, free expression and the free press made by the” BTLR Report is open for signature at the House of Commons website. Please sign it!

What Does Vavilov Stand For?

This post is co-written with Leonid Sirota.

As we previously noted in a joint post on Canada (Minister of Citizenship and Immigration) v Vavilov, 2019 SCC 65, that decision leaves open the question whether reasonableness review, as explained in the majority reasons, tends toward deference or vigilance, and so whether it will be more rigorous than pre-Vavilov reasonableness. After all, Vavilov begins by saying that its application of the reasonableness standard is based on a principle of judicial restraint, one which “demonstrates a respect for the distinct role of administrative decision-makers” (Vavilov, at para 13). Yet in the same breath the majority insists that reasonableness “remains a robust form of review” (Vavilov, at para 13). It then adds that the reasonableness standard is strong enough to guard against threats to the Rule of Law—yet not so strong as to revert to a form of jurisdictional review (Vavilov, at para 67).

Because of these statements, it is not surprising that some suggest that Vavilov is more robust than restrained, while others view it as “inherently deferential”. At first blush, these different takes on Vavilov could be argued to reflect confusion at the heart of the decision. And Vavilov’s rhetoric is indeed confusing. But an optimistic interpretation of the majority’s reasons might be that they speak to the great variety of cases to which they apply. In some, review will be more constrained; in others, it will be more rigorous. It will be the task of lower courts to parse the Vavilov judgment to determine which circumstances call for which application. But it is not clear that Vavilov prescribes an approach to judicial review that is uniformly more or less restrained.

While it is too early to draw any trend lines, the lower courts have had a chance to weigh in on this question, and they too are divided. But taking the cases together, they might support the conclusion that Vavilov is more contextual than categorical. On one hand, some cases have put forward more interventionist readings of Vavilov. One of us wrote here about Canadian National Railway Company v Richardson International Limited, 2020 FCA 20. There Nadon JA applied the appellate correctness standard, but he added that had he applied reasonableness review, he would have found the decision unreasonable. Nadon JA faulted the Canadian Transportation Agency for failing to take account of statutory context by focusing too much on the text of the relevant statutory provision—and said this would have been just as much of a problem on reasonableness review. Similarly, in Farrier c Canada (Procureur général), 2020 CAF 25, Gauthier JA explained that while she might have found a decision of the Appeal Division of the Parole Board of Canada reasonable under Dunsmuir and its progeny, under Vavilov, the story was different (Farrier, at paras 12, 19). The failure of the administrative decision-maker to provide reasons on some key legal elements of the decision was fatal.

By their own admission, these cases take a harder look at the administrative decisions under review than one would have expected prior to Vavilov, especially in how they scrutinize the administrative decision-makers’ reasons. This seems fully consistent with Vavilov’s “reasons first” approach to judicial review (Vavilov, at para 84) and its clear rejection of the practice of judicial supplementation of reasons (Vavilov, at paras 96-97). Coupled with these changes, Vavilov introduces what one might call a “legal hard look review”. There is now an expectation that administrators will reason with reference to their enabling statutes and take account these statutes’ text, context, and purpose (Vavilov, at para 120). Their omissions in this regard can only be tolerated if they are minor (Vavilov, at para 122). But, as Richardson holds, a failure to justify a decision in relation to the statute at all will be fatal.

By contrast, some decisions in Ontario’s Divisional Court fail to see a meaningful difference between Vavilov and the previous judicial review regime. In Radzevicius v Workplace Safety and Insurance Appeals Tribunal, 2020 ONSC 319 , for example, Swinton J asserts that “Vavilov does not constitute a significant change in the law of judicial review with respect to the review of the reasons of administrative tribunals” (Radzevicius at para 57). She holds that, because there was no “fundamental flaw or gap in the Tribunal’s analysis”, the decision was reasonable (Radzevicius, at para 59). And in Correa v Ontario Civilian Police Commission, 2020 ONSC 133, Swinton J (writing for a differently composed panel) was similarly disposed, insisting that Vavilov did not impose a “more robust reasonableness review” (Correa, at para 54).

These cases divide on a basic question: is Vavilov reasonableness review more robust than what preceded it? The temptation is to fall on one or another side of this question, but the truth might be a bit more nuanced. The decisions we have just discussed suggest that, while Vavilov may impose more robust review in some circumstances, it is also possible that, in other cases, Vavilov will, indeed, not change the degree of deference.

In Richardson and Farrier, at issue were questions of legal interpretation: questions that required the decision-maker to engage with the enabling statute to determine the meaning, in context, of relevant provisions. A failure to engage with key elements of the statute, insofar as such a lack of consideration may change the result, is not reasonable, according to Vavilov (Vavilov, at para 122). It is probably fair to say that, at least when it comes to pure questions of statutory interpretation, reasonableness review may now take on a more interventionist flavour, particularly where decision-makers are not simply applying a statutory provision to facts but are actually attempting to determine the meaning of particular provisions. When decision-makers are interpreting a statute, Vavilov demands more of them than the cases it supersedes.

Radzevicius and Correa, by contrast, presented issues of mixed fact and law. Decision-makers having applied legal standards to particular facts and evidence; they did not fail to take account of relevant legal provisions or fundamentally misunderstand them. Vavilov says that while the evidence acts as a constraint on the decision-maker, courts must refrain from reweighing evidence or holding decision-makers to a high evidentiary standard (Vavilov, at para 125). This injunction is consistent with Vavilov’s judicial restraint theme. When evidence is more central to the disposition of the case, we might expect review to be more deferential. Whatever else it did, Vavilov did not—for better or for worse—bring back the concept of jurisdictional fact.

This division finds some theoretical support in the literature, specifically Jeffrey Pojanowski’s recent paper advocating for a “neoclassical” approach to administrative law (which one of us reviewed here). Under the neoclassical approach, courts take a harder look at agency legal interpretations while respecting agency space to maneuver on policy or evidentiary matters (883). Neoclassicism pays attention to what both a particular decision-maker’s enabling statute and general legislation, such as the American Administrative Procedure Act indicate about the intensity of the review to which the decision-maker is subject. This approach is to be contrasted with, among others, “administrative supremacy”, which advocates across-the-board deference on all questions of law, reducing the rule of law to a “thin residue” around the margins of delegated power (869).

Vavilov, on the understanding expounded in this post, lends itself to a neoclassical interpretation. On one hand, it asks decision-makers to specifically reason in relation to the limitations on their power, most notably their enabling statute (Vavilov, at paras 108-110, 120). It introduces new requirements to engage with the text, context, and purpose of the statute (Vavilov at para 118 et seq). All of these requirements are rooted in the centrality of the enabling statute, and the role of courts to interpret that statute to decide on the intensity of review. But on the other hand, Vavilov largely incorporates existing law in asking courts to stay their hand when it comes to the evidence before a decision-maker, and the way it might have been assessed (see Khosa, at para 61). In part, this can be justified as a dutiful reflection by the courts that the decision-maker was (1) the initial merits decider and (2) Parliament’s chosen delegate, established to be the merits decider. This division is therefore rooted in a plausible understanding of the respective roles of courts and delegated decision-makers.

Some caveats are in order. First, the distinction between questions of law and questions of evidence will not necessarily be perfectly neat. Indeed, it is true that sometimes, on legal questions, a decision-maker will have a wide margin in which to operate because of the words of a statutory grant of authority (Vavilov, at para 110). As a result, the distinction we draw here might not be helpful in every case. Our point is simply that it may help explain how courts have thus far treated Vavilov. And second, we do not know whether this distinction is really what drove the courts’ reasoning in these cases. Or was their reasoning, instead, primarily a function of individual or institutional views on judicial review, which are bound to influence judges as they work to make sense of equivocal guidance from the Supreme Court? It will be interesting to see, for example, how the Federal Court of Appeal treats more fact-bound cases and, conversely, how the Divisional Court will approach those where statutory interpretation is at the forefront.

That said, if there is one thing that is clear about Vavilov, it is that the various constraints that operate to limit the space within which a decision-maker can maneuver are supposed to be sensitive to context. As the relevant facts and applicable law vary, so different constraints come to the fore. The constraints that apply in a given case lead to more or less interventionist review. One of us suggested, in a contribution to the symposium on the tenth anniversary of Dunsmuir, that the administrative law framework that should replace the one that built on Dunsmuir (or on its ruins) ought to “abandon the pursuit, or the pretense, of across-the-board deference” in favour of greater sensitivity “to the circumstances of particular cases … As these circumstances vary, so must the applicable rules.” Ostensibly, Vavilov instead doubles down a one-size-fits-most reasonableness standard of review. But it may be that, in practice, it makes sufficient room for a more nuanced approach.

In the short term, this might lead to more confusion. In the long run, however, it may prove a more fruitful way of developing the Canadian law of judicial review. For now, it is for the lower courts to work out the precise circumstances in which more or less deference is due to administrative decision-makers. As a result, confident broad judgments about Vavilov’s true import are probably premature.

Richardson: Rigorous Vavilov Review

In one of the Federal Court of Appeal’s post-Vavilov cases, CNR v Richardson, the Court (per Nadon JA) demonstrates that Vavilov review, on substantive questions of law, will not be inattentive or subordinate to administrative discretion. Indeed, while some suggest that Vavilovian review is “inherently deferential,” I see the matter quite differently:  Richardson shows how Vavilov review puts the court in the proper position to rigorously enforce the statutory boundaries of administrative decision-making, particularly where decision-makers fail to engage with elements of the statute at all.

Richardson is an agri-food business that owns and operates 54 grain elevators [3]. CN and CP serve a number of Richardson’s elevators [3].  Meanwhile, the Canadian National Railway Company (CN) and the Canadian Pacific Railway Company (CP) connect railway networks at Scotford, Alberta [5], though the mainlines of each do not connect at Scotford. Around this connection (about 30km) lays Richardson’s Lamont elevator, located on CN’s main line [6].

Richardson filed an application before the Canada Transportation Agency, asking that the Scotford site be deemed an “interchange” and that Richardson traffic be transferred for “interswitching” between Scotford and Lamont elevator. Under the relevant statute (the Canada Transportation Act), an interchange “means a place where the line of one railway company connects with the line of another railway company and where loaded or empty cars may be stored until delivered or received by the other railway company” (s.111). Meanwhile, “interswitch” “means to transfer traffic from the lines of one railway company to the lines of another railway company” (s.111). The goal of interswitching is to ensure that shippers with only one choice of railway have “fair and reasonable access to the rail system at a reasonable rate.” Obviously, this would benefit Richardson.

The Agency first concluded that CN and CP operated an interchange, under the statute, at Scotford. While the Agency noted CN’s argument that its main line did not connect with CP’s main line at Scotford, it ultimately held that s.111 of the statute did not make a distinction between the type of railway line required to make the connection [11]. Rather, under the statute, a railway line is defined broadly [11]. Further, s.140 of the statute outlines exemptions for what a “railway line” does not include. To the Agency, if Parliament wanted to limit interchanges to areas where main lines connected, it could have include a similar exclusion in s.111 of the statute [11].

On judicial review, CN argued that the Agency erred in interpreting the definition of “interchange.” The first question for Nadon JA was the standard of review. He concluded that while, pre-Vavilov, the standard was reasonableness, the case came before the Court on a statutory right of appeal, which meant that the standard of correctness applied [42-44]. But Nadon JA went on to conclude in obiter that even though the standard of review was correctness, “[u]nder the previous standard of reasonableness, I would have had no hesitation concluding that the Agency’s interpretation was unreasonable because it failed to consider both context and the legislative scheme as a whole” [46]. Moreover, the Agency misapplied a principle of statutory interpretation—the so-called “implied exclusion rule” adopted by the Agency was inconsistent with Supreme Court precedent  (see Green, at para 37). As such, Justice Nadon remitted to the Agency to receive its view about the interpretive matter, especially in light of the new correctness standard.

Nadon JA’s analysis is tightly connected to Vavilov, and demonstrates how courts should apply Vavilov in light of defective statutory reasoning. As Vavilov notes, at para 108, the governing statute is the most “salient” aspect of the context bearing on a decision-maker. As such, this most salient aspect must be rigorously enforced against the decision-maker. But a court cannot do so in absence of reasons from the decision-maker engaging with the statute.  Indeed, the reasons must demonstrate some engagement with this and other constraints (Vavilov, at para 120). It is for this reason that decision-makers are required to interpret the law in concert with its text, context, and purpose (see Vavilov, at paras 118, and 120). While the Agency is not required to engage in a “formalistic analysis” (Vavilov, at para 119), and not all errors will be material (Vavilov, at para 122), a failure to engage with purpose at all as in Richardson must be considered fatal, if the governing statutory scheme is a real constraint on administrative decision-making. More specifically, purpose and the overall context must always be considered in run-of-the-mill statutory interpretation cases, and a failure to do so is fatal in that context (see ATCO Gas & Pipelines, 2006 SCC 4 at para 48), as Vavilov says it can be in the administrative context. And what’s more, a failure to use the proper principles of interpretation is a failure of reasoning that Vavilov says is impermissible, because a decision-maker’s interpretation of a statutory provision must be consistent with the text, context, and purpose of a provision, and because “the usual principles of statutory interpretation apply equally when an administrative decision maker interprets a provision” (Vavilov, at para 120).

Richardson, then, is on solid ground. The Agency’s failure to consider the overall statutory context means that it was not “alive” to an essential element of interpretation (Vavilov, at para 120). A failure to engage with the statute at all cannot, on any understanding of the term, be reasonable. And the Agency’s use of the implied exclusion rule, without considering the broader context and purpose, clearly runs counter to the Supreme Court’s comments in Green about the implied exclusion rule. The failure to properly engage with the statute in its entirety, without using the proper tools of interpretation, is a Vavilovian error.

There is a broader point of principle here. As Vavilov implicitly holds, it is not formalistic to expect decision-makers, who share in the enterprise of law-making, to actually do the task properly. Indeed, proponents of deference cannot say two things at once: they cannot insist that decision-makers are contributors to law-making, but then grant decision-makers the ability to engage in reasoning that does not engage with the most obvious and natural limitation on administrative decision-making (Vavilov, at para 109). Either decision-makers are shared partners in law-making or they are not.

It is true that reasons are the focus of Vavilov, and so it is methodologically deferential to look at those reasons first. But this does not lessen the rigorousness of Vavilovian review, at least as exemplified by Nadon JA in Richardson. Indeed, the reasons are merely the window into the application of the constraints on the decision-maker. They are not an invitation for courts to sit back—reasons require responsive engagement by the Court, in relation to the application of the principles of interpretation.

While Nadon JA’s remarks in Richardson are obiter, they are a good example of the promise of Vavilov: rigorous reasonableness review that is focused on the statute and in ensuring that administrative decision-makers engage with the statute.

 

Tout nouveau, tout beau?

Ce que dit, et ce que ne dit pas, l’arrêt Vavilov, pour nos lecteurs francophones

Ce billet est co-rédigé avec Mark Mancini

L’arrêt Canada (Ministre de la Citoyenneté et de l’Immigration) c Vavilov, 2019 CSC 65 de la Cour suprême a fait l’objet de nombreux commentaires, tant sur ce blogue qu’ailleurs, – mais dans la langue de Laskin, pas celle de Beetz. Nous nous proposons donc de combler ce vide. Ce billet ne saurait reprendre les analyses et les critiques détaillées que nous avons tous deux déjà publiées (dont la liste suit ci-dessous) et celles, peut-être, encore à venir. Il se limite plutôt, d’une part, à offrir à nos lecteurs francophones un résumé des points saillants de l’arrêt et, de l’autre, à attirer leur attention sur les enjeux que risque de soulever la mise en œuvre de celui-ci par les tribunaux.

Ainsi qu’elle l’avait annoncé dans son jugement accordant l’autorisation de pourvoi, la Cour suprême profite de l’affaire Vavilov pour ajuster le cadre d’analyse employé par les tribunaux lorsqu’ils révisent une décision administrative sur le fond. Si les normes de contrôle disponibles demeurent celles que les tribunaux canadiens ont appliquées depuis l’arrêt Dunsmuir c Nouveau-Brunswick, 2008 CSC 9, [2008] 1 RCS 190, et que la présomption de l’application de la norme de la décision raisonnable demeure en vigueur, tant les fondements théoriques de ce cadre d’analyse que les circonstances où la présomption est repoussée sont révisées. De plus, la Cour fournit des explications étoffées sur la façon d’appliquer la norme de contrôle de la décision raisonnable, qui seront sans doute un enseignement tout aussi important, et probablement plus difficile à appliquer, de cet arrêt.


Le principe qui guide le choix de la norme de contrôle appliquée lors de la révision d’une décision administrative est celui voulant que cette norme doit « refléter l’intention du législateur sur le rôle de la cour de révision, sauf dans les cas où la primauté du droit empêche de donner effet à cette intention » [23]. Selon la Cour, cela signifie généralement que, « [s]i le législateur a constitué un décideur administratif dans le but précis d’administrer un régime législatif […] on peut aisément présumer que le législateur a voulu que celui‑ci puisse fonctionner en faisant le moins possible l’objet d’une intervention judiciaire » [24]. Il s’ensuit que c’est la norme de contrôle empreinte de déférence, soit celle de la décision raisonnable, qui s’applique – en principe.

Il faut bien noter que c’est le seul choix du législateur qui dicte cette conclusion. L’expertise réelle ou présumée du décideur administratif n’y est pour rien, à la différence de ce qui a pu être le cas dans la jurisprudence (dont l’arrêt Edmonton (Ville) c Edmonton East (Capilano) Shopping Centres Ltd, 2016 CSC 47, [2016] 2 RCS 293 est un exemple particulièrement frappant). La notion d’expertise n’est pas tout à fait reléguée aux oubliettes – nous y reviendrons –, mais son exclusion de l’analyse quant choix de la norme de contrôle a des conséquences importantes, et pourrait en avoir d’autres, non moins significatives. Nous y reviendrons aussi.

La présomption voulant que la norme de contrôle d’une décision administrative soit celle de la décision raisonnable est repoussée dans deux cas. Le premier est celui où le législateur a lui-même indiqué qu’une autre norme de contrôle est applicable. Il peut le faire en légiférant directement sur le sujet. Il peut aussi, cependant, le faire en créant un droit d’appel – avec ou sans autorisation – à une cour de justice. Lorsqu’elle siège en appel d’une décision administrative, c’est la norme de contrôle qui s’appliquerait à une question équivalente dans un appel d’une décision judiciaire que la cour doit appliquer. Ainsi, « elle se prononcera sur des questions de droit, touchant notamment à l’interprétation législative et à la portée de la compétence du décideur, selon la norme de la décision correcte » [37]. Il s’agit là d’un changement important par rapport à la jurisprudence précédente qui, suivant l’arrêt Pezim c ColombieBritannique (Superintendent of Brokers), [1994] 2 RCS 557, recourait généralement, même en appel, à la norme de contrôle de révision judiciaire, en raison notamment de l’expertise supposée des décideurs administratifs. (Notons, cependant, « que ce ne sont pas toutes les dispositions législatives envisageant la possibilité qu’une cour de justice puisse contrôler une décision administrative qui confèrent dans les faits un droit d’appel » [51]. En particulier, l’arrêt Canada (Citoyenneté et Immigration) c Khosa, 2009 CSC 12, [2009] 1 RCS 339 et son interprétation, qui nous semble erronée, de la Loi sur les cours fédérales, ne semblent pas affectés par Vavilov.)

Le second cas où la présomption de l’application de la norme de la décision raisonnable est repoussée est celui où son application serait contraire à la primauté du droit. Vavilov enseigne que celle-ci exige une réponse correcte, et non seulement raisonnable, à trois types de questions. Il s’agit, en premier lieu, de questions de validité constitutionnelle; en deuxième lieu, de « questions de droit générales d’une importance capitale pour le système juridique dans son ensemble » [53]; et, en troisième lieu, de celles concernant « la délimitation des compétences respectives d’organismes administratifs » [63]. D’autres types de questions pourraient, en principe, s’ajouter à cette liste, mais la Cour semble plutôt sceptique à ce sujet.

Trois observations s’imposent ici. Premièrement, s’agissant de questions constitutionnelles, Vavilov ne remet pas en cause – à première vue en tout cas – l’arrêt Doré c Barreau du Québec, 2012 CSC 12, [2012] 1 RCS 395. La Cour souligne expressément qu’elle ne se prononce pas sur la validité du cadre d’analyse qui y a été établi. Deuxièmement, s’agissant de « questions d’une importance capitale », cette catégorie se trouve possiblement élargie en comparaison avec le cadre d’analyse de l’arrêt Dunsmuir, puisqu’elle ne dépend plus d’une évluation de l’expertise relative du tribunal et du décideur administratif. Troisièmement, la catégorie de « véritables questions de compétence », retenue dans Dunsmuir et préservée, en ne serait-ce qu’en théorie, dans la jurisprudence subséquente, est abolie par Vavilov, du moins au stade du choix de la norme de contrôle.


Ces ajustements au choix de la norme de contrôle apportés, la Cour se tourne vers la norme de la décision raisonnable. Elle explique que « le contrôle selon la norme de la décision raisonnable a pour point de départ la retenue judiciaire et le respect du rôle distinct des décideurs administratifs » [75]. Ce contrôle vise néanmoins à s’assurer que le décideur administratif tienne compte des « contraintes juridiques et factuelles auxquelles [il] est assujetti » [85] et qu’il explique sa décision à ceux et celles qu’elle affecte.

Les motifs du décideur administratif occupent donc une importance centrale dans le contrôle judiciaire – et ce, même si la Cour suprême reconnaît qu’un décideur n’est pas toujours tenu de les rédiger. C’est le raisonnement du décideur administratif, tel que représenté dans les motifs, qui fait l’objet d’examen :

Une cour de justice qui applique la norme de contrôle de la décision raisonnable ne se demande donc pas quelle décision elle aurait rendue à la place du décideur administratif, ne tente pas de prendre en compte l’ « éventail » des conclusions qu’aurait pu tirer le décideur, ne se livre pas à une analyse de novo, et ne cherche pas à déterminer la solution « correcte » au problème. [83]

La cour de révision ne doit pas, non plus, « élabore[r] ses propres motifs pour appuyer la décision administrative » ou encore « faire abstraction du fondement erroné de la décision et […] y substituer sa propre justification du résultat ». [96] Cependant, les motifs ne sont pas tenus à la perfection et peuvent, le cas échéant, être lus à la lumière du dossier. Les motifs peuvent également permettre au décideur de démontrer son expertise et d’ainsi justifier « un résultat qui semble déroutant ou contre‑intuitif à première vue » comme étant « néanmoins conforme aux objets et aux réalités pratiques du régime administratif en cause » [93].

Appliquant la norme de la décision raisonnable, la cour de révision s’intéresse donc à la fois au raisonnement du décideur et au résultat auquel celui-ci a abouti. Les deux doivent être justifiables et justifiés. La Cour suprême propose une liste, qui se veut non-exhaustive, « de questions qui peuvent révéler qu’une décision est déraisonnable » [101]. Certaines concernent la cohérence du raisonnement du décideur administratif. Une décision irrationnelle, entachée de paralogismes, dont « la conclusion […] ne peut prendre sa source dans l’analyse effectuée » [103] ou celle dont « il est impossible de comprendre, lorsqu’on lit les motifs en corrélation avec le dossier, le raisonnement […] sur un point central » [103] doit être traitée comme déraisonnable.

Tel est aussi le cas d’une décision qui ne tient pas compte du contexte juridique et factuel dans lequel elle est rendue. La Cour souligne que

le régime législatif applicable est probablement l’aspect le plus important du contexte juridique d’une décision donnée. Le fait que les décideurs administratifs participent, avec les cours de justice, à l’élaboration du contenu précis des régimes administratifs qu’ils administrent, ne devrait pas être interprété comme une licence accordée aux décideurs administratifs pour ignorer ou réécrire les lois adoptées par le Parlement et les législatures provinciales. [108]

D’une part, même lorsque le décideur administratif jouit d’un pouvoir discrétionnaire, « tout exercice d’un [tel] pouvoir […] doit être conforme aux fins pour lesquelles il a été accordé » [108]. De l’autre, « un organisme administratif ne saurait exercer un pouvoir qui ne lui a pas été délégué ». [109] La porté du pouvoir délégué ou l’étendue des raisons de cette délégation varie selon le texte législatif applicable. Le contrôle en vertu de la norme de la décision raisonnable exige donc de la cour de révision « de déterminer si […] le décideur a justifié convenablement son interprétation de la loi à la lumière du contexte. Évidemment, il sera impossible au décideur administratif de justifier une décision qui excède les limites fixées par les dispositions législatives qu’il interprète ». [110]

La marge de manœuvre du décideur administratif dépend, en outre, des autres lois ou règles du droit prétorien qui peuvent s’appliquer à la décision. La décision administrative doit, notamment, tenir compte des règles d’interprétation législative, sans pour autant forcément « procéder à une interprétation formaliste de la loi » [119]. Le décideur administratif peut tenir compte de ses connaissances et de son expertise spécialisées, mais « il [lui] incombe […] de démontrer dans ses motifs qu’il était conscient [des] éléments essentiels » [120] de l’interprétation législative, et il ne lui est pas loisible d’ « adopter une interprétation qu’il sait de moindre qualité — mais plausible — simplement parce que cette interprétation paraît possible et opportune » [121].

Par ailleurs, une décision administrative doit aussi se justifier au regard de la preuve, des arguments des parties et de la pratique administrative. Elle doit aussi refléter, le cas échéant, son importance pour la personne visée : « Lorsque la décision a des répercussions sévères sur les droits et intérêts de l’individu visé, les motifs fournis à ce dernier doivent refléter ces enjeux. […] Cela vaut notamment pour les décisions dont les conséquences menacent la vie, la liberté, la dignité ou les moyens de subsistance d’un individu » [133].

Un dernier enseignement en matière de l’application de la norme de la décision raisonnable sur lequel nous voudrions attirer l’attention du lecteur concerne les réparations que peut accorder une cour de révision. La Cour suprême explique que « lorsque la décision contrôlée selon la norme de la décision raisonnable ne peut être confirmée, il conviendra le plus souvent de renvoyer l’affaire au décideur pour qu’il revoie la décision, mais à la lumière cette fois des motifs donnés par la cour ». [141] Cependant, et il s’agit, dans une certaine mesure, d’une nouveauté, la Cour précise qu’ « il y a des situations limitées » [142] où la cour de révision doit elle-même trancher le différend, pour éviter de le prolonger inutilement. C’est notamment le cas lorsqu’une seule réponse est possible a une question d’interprétation, mais d’autres facteurs, y compris ceux concernant les coûts, tant pour les parties que pour l’administration et le système de justice, doivent aussi être pris en compte.


L’arrêt Vavilov promet – pas pour la première fois en droit administratif canadien – « d’apporter une cohérence et une prévisibilité accrues à ce domaine du droit ». [10] Cette promesse sera-t-elle tenue? À certains égards, les enseignements de la Cour suprême sont prometteurs. Notamment, la nouvelle approche au choix de la norme de contrôle, qu’on soit ou non d’accord avec la présomption du choix de la norme de la décision raisonnable ou encore avec l’abolition de la catégorie de questions de compétence, promet du moins une certaine simplification par rapport à l’état du droit avant Vavilov. L’insistance de la Cour sur l’importance des motifs et du respect du cadre législatif par les décideurs administratifs est elle aussi plus que bienvenue.

Plusieurs questions importantes restent toutefois sans réponse. Les cours de révision, et éventuellement la Cour suprême elle-même, devront y répondre pour que l’on puisse véritablement affirmer que le droit administratif canadien est simple est prévisible. En voici quelques unes.

Quelle sera la portée réelle des catégories de questions où la primauté du droit exige l’application de la norme de la décision correcte? En particulier, quel avenir réserve la Cour à l’arrêt Doré?

Comme nous l’avons souligné ci-dessus, l’arrêt Vavilov semble élargir quelque peu la catégorie de questions « d’une importance capitale pour le système juridique », en raison de l’abolition de la référence à l’expertise dans sa délimitation. Or, si la Cour résume la jurisprudence existante à ce sujet et dit que celle-ci « continue de s’appliquer essentiellement telle quelle » [143], ce résumé ne fournit que des exemples, et non de véritables lignes directrices. L’incertitude risque de persister à ce sujet.

Plus grave encore, mais peut-être susceptible d’une résolution plus rapide, est l’incertitude quant à l’avenir du cadre d’analyse posé dans l’arrêt Doré et raffiné ou modifié dans École secondaire Loyola c Québec (Procureur général), 2015 CSC 12, [2015] 1 RCS 613 et Law Society of British Columbia c Trinity Western University, 2018 CSC 32, [2018] 2 R.C.S. 293. La Cour, nous l’avons déjà dit, se garde de se prononcer explicitement à ce sujet. Pourtant, les fondements de cette jurisprudence, qui repose en bonne partie sinon entièrement sur la volonté de respecter l’expertise – réelle ou supposée – des décideurs administratifs, nous semblent incompatibles avec l’exclusion de l’expertise de l’analyse quant au choix de la norme de contrôle dans Vavilov. De plus, nous sommes sceptiques face à l’idée que le législateur puisse dicter, implicitement ou même explicitement, le choix de la norme de contrôle en matière constitutionnelle, qu’il s’agisse de questions de validité ou des celles concernant la constitutionnalité de décisions particulières. La Cour suprême le dit fort bien dans Vavilov : « si un législateur peut choisir les pouvoirs à déléguer à un organisme administratif, il ne peut déléguer des pouvoirs dont la Constitution ne l’investit pas. Le pouvoir constitutionnel d’agir doit comporter des limites définies et uniformes, ce qui commande l’application de la norme de la décision correcte » [56].

Les questions de compétence sont-elles véritablement à oublier?

La catégorie de « véritables questions de compétence » est écartée de l’analyse quant au choix de la norme de contrôle. Pourtant, en affirmant que « certaines questions touchant à la portée du pouvoir d’un décideur […] ne sauraient commander qu’une seule interprétation », et qu’ « [é]videmment, il sera impossible au décideur administratif de justifier une décision qui excède les limites fixées par les dispositions législatives qu’il interprète », [110] la Cour semble tout simplement utiliser une nouvelle étiquette pour la décrire. Par ailleurs, les tribunaux pourraient être appelés à décider une question en est une de compétence en disposant d’appels autorisés par des dispositions législatives qui y font référence.

Comment la norme de la décision raisonnable sera-t-elle appliquée en l’absence de motivation adéquate par le décideur administratif?

Si l’on peut se réjouir du fait que la Cour suprême semble souhaiter mettre un frein à la tendance, qui s’est parfois manifestée dans la jurisprudence, de l’écriture rétroactive des motifs de décision administrative par les cours de révision, on peut se demander jusqu’où sa détermination ira en pratique. La Cour insiste, d’une part, pour dire qu’une décision administrative qui doit être motivée mais ne l’est pas ou ne l’est pas adéquatement sera déraisonnable, mais, d’autre part, elle souligne « qu’une cour de révision doit examiner le dossier dans son ensemble pour comprendre la décision et qu’elle découvrira alors souvent une justification claire pour la décision » [137]. L’équilibre entre ces deux exigences ne nous semble pas évident à trouver.

De la déférence à l’égard du décideur administratif et de la vigilance quant au respect du cadre législatif, laquelle va l’emporter de l’application de la norme de la décision raisonnable?

La Cour offre, à ce sujet, des enseignements qui peuvent sembler contradictoires. Elle affirme, notamment, dans un seul et même court paragraphe, que « [l]e contrôle selon la norme de la décision raisonnable […] tire son origine du principe de la retenue judiciaire », mais aussi que « [c]e type de contrôle demeure rigoureux ». [13] Comment la cour de révision s’y prendra-t-elle pour exercer son pouvoir avec retenue et vigueur à la fois? Comment va-t-elle déterminer si un décideur administratif a respecté les contraintes que la loi lui imposait sans pour autant tenter de délimiter l’ « évantail » des solutions possibles, ou encore vérifier s’il a respecté les principes d’interprétation législative tout en gardant à l’esprit que « La ‘‘justice administrative’’ ne ressemble pas toujours à la ‘‘justice judiciaire’’ » [92]?

Le fondement théorique de l’arrêt Vavilov, soit le respect de la volonté du législateur (circonscrit par le principe de la primauté du droit, mais déterminant dans les limites que celui-ci impose), ne permet pas de résoudre cette tension. S’il est vrai que le législateur confie l’application et donc la première interprétation de la loi au décideur administratif, c’est aussi le législateur qui choisir de limiter le pouvoir discrétionnaire de ce dernier par le texte de loi qu’il adopte. Il faudra donc voir comment les tribunaux, y compris la Cour suprême elle-même, appliqueront la norme de contrôle de la décision raisonnable, et s’ils parviendront à résoudre les tensions présentes dans les motifs de la Cour. Ce n’est qu’en cas de succès, qui n’est pas acquis d’avance, que l’on pourra affirmer que l’arrêt Vavilov a véritablement réglé les problèmes de cohérence et de prévisibilité du droit administratif auxquels la Cour suprême s’y attaquait.


L’arrêt Vavilov sera, évidemment, un jalon important dans le développement du droit administratif canadien. Cependant, ses silences et ses contradictions pourraient s’avérer tout aussi importants que ses enseignements. Aussi important ce jalon soit-il, il est loin de marquer la fin du parcours souvent tortueux de ce domaine du droit.


Voici la liste, mentionnée ci-dessus, de billets que nous avons publiés sur l’arrêt Vavilov et ses conséquences, en ordre chronologique:

Chevron on 2

The illogic of the Supreme Court of Canada’s approach to deference to administrative interpretations of law

Readers with some salsa experience will probably know that, while most of the world dances it “on 1”, in New York it is danced “on 2”. The steps and moves are more or less the same, but the sequence is different. Another dance that can be varied in this way, as we learn from the Supreme Court’s recent decision in Canada (Minister of Citizenship and Immigration) v Vavilov, 2019 SCC 65, is the notorious Chevron two-step. As with salsa, one can prefer one style or the other. But, for what it’s worth, I find Vavilov’s “on 2” version of Chevron to be rather offbeat.


In Chevron USA Inc v Natural Resources Defense Council, Inc, 467 US 837 (1984), the US Supeme Court explained how courts were to review administrative decision-makers’ interpretations of what in Canada are sometimes called their “home statutes”:

When a court reviews an agency’s construction of the statute which it administers, it is confronted with two questions. First, always, is the question whether Congress has directly spoken to the precise question at issue. If the intent of Congress is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress. If, however, the court determines Congress has not directly addressed the precise question at issue, the court does not simply impose its own construction on the statute, as would be necessary in the absence of an administrative interpretation. Rather, if the statute is silent or ambiguous with respect to the specific issue, the question for the court is whether the agency’s answer is based on a permissible construction of the statute. (842-43; footnotes omitted)

The first step, in other words, is to determine whether the statute is so vague or ambiguous as to require an exercise of interpretive discretion by the administrative decision-maker. The second step, taken if―and only if―the statute does call for such an exercise of discretion, is to review the administrative interpretation for reasonableness, and defer to it if it is not unreasonable.

There are some exceptions to this two-step analysis. For one thing, under United States v Mead Corp, 533 US 218 (2001), courts ask whether the administrative agency was meant to conclusively determine questions of law in the first place. This is sometimes known as “Chevron step zero”. For another, following FDA v Brown & Williamson Tobacco Corp, 529 US 120 (2000), certain questions are seen as too important for their determination to have been delegated to administrative agencies implicitly; nothing short of explicit Congressional command will trigger deference. But, at least where the administrative decision-maker is seen as authorized to make legal determinations, Chevron dictates ― for now anyway ― the normal approach.

Or, if you prefer seeing and hearing instead of reading, here’s how NYU students explained it a few years ago:


Now, compare this to the Vavilov framework. It begins with a fairly close equivalent to “Chevron step zero”. In cases where the legislature wanted the courts, and not administrative tribunals, to decide legal questions, whether by explicitly providing for correctness review or by creating an appeal from from the tribunal to a court, the courts must not defer. Nor will there be deference on (some) constitutional questions and “general questions of law that are ‘of central importance to the legal system as a whole'” [58, quoting Dunsmuir v. New Brunswick, 2008 SCC 9, [2008] 1 SCR 190 [62]]. This is somewhat analogous to the “important questions” exception in the United States, although Canadian “questions of central importance” may well be different from the American “important questions”. (I don’t think, for instance, that under Vavilov it is enough for a question to be “of deep economic and political significance [and] central to [a] statutory scheme”: King v Burwell (2015) (internal quotation omitted).)

But then, Chevron‘s two main steps are reversed. Subject to the legislative assignment and central questions exceptions applying, Vavilov says courts are to defer to administrative interpretations of law:

Where a legislature has created an administrative decision maker for the specific purpose of administering a statutory scheme, it must be presumed that the legislature also intended that decision maker to be able to fulfill its mandate and interpret the law as applicable to all issues that come before it. Where a legislature has not explicitly prescribed that a court is to have a role in reviewing the decisions of that decision maker, it can safely be assumed that the legislature intended the administrative decision maker to function with a minimum of judicial interference. [24]

This is, more or less, Chevron‘s step two. At this stage, no factor other than the existence of the administrative decision-maker, the absence of a legislative indication that courts must nevertheless be involved, and the non-centrality of the question at issue are relevant.

But then, Vavilov seems to suggest that, once it embarks on reasonableness review, the court needs to examine the statute at issue more closely ― to engage what co-blogger Mark Mancini has described as a “legal ‘hard look’ review”, including to determine whether there is actually the sort of ambiguity that, under Chevron, justifies deference to the administrative interpretation. Vavilov stresses that “while an administrative body may have considerable discretion in making a particular decision, that decision must ultimately comply ‘with the rationale and purview of the statutory scheme under which it is adopted'” [108, quoting Catalyst Paper Corp v North Cowichan (District), 2012 SCC 2, [2012] 1 SCR 5, [15]] and, further, “with any more specific constraints imposed by the governing legislative scheme”. [108] Crucially, Vavilov insists that

If a legislature wishes to precisely circumscribe an administrative decision maker’s power in some respect, it can do so by using precise and narrow language and delineating the power in detail, thereby tightly constraining the decision maker’s ability to interpret the provision. Conversely, where the legislature chooses to use broad, open-ended or highly qualitative language … it clearly contemplates that the decision maker is to have greater flexibility in interpreting the meaning of such language. … [C]ertain questions relating to the scope of a decision maker’s authority may support more than one interpretation, while other questions may support only one, depending upon the text by which the statutory grant of authority is made. [110]

This, by my lights, is Chevron‘s step one. In some cases, the Supreme Court says, the legislature leaves the administrative decision-maker with the latitude to choose among competing possible interpretations. But not always. To quote Chevron again, “[i]f the intent of [the legislature] is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of [the legislature]”.

I should note that this might not be the only way to read Vavilov. Paul Daly, for example, is quite skeptical of “intrusive reasonableness review” that would occur if courts take too seriously the admonition about there being, sometimes, only one interpretation of administrative decision-maker’s grant of authority. But, as Mark shows, this is certainly a plausible, and at least arguably the better reading of Vavilov. I may return to the debate between these readings in a future post. For now, I will assume that the one outlined above is at least a real possibility.


As already mentioned, this reversal of the “Chevron two-step” makes no sense to me. I find it odd to say that reviewing courts must start from the position that “respect for [the] institutional design choices made by the legislature” in setting up administrative tribunals “requires a reviewing court to adopt a posture of restraint on review”, [24] but then insist that respect for legislative choices also requires the courts to be vigilant in case these choices leave only one permissible interpretation. The view, endorsed in Dunsmuir, that deferential judicial review reflects the inherent vagueness of legal language, was empirically wrong (and indeed implausible, as I argued here), but coherent. The recognition in Vavilov that statutory language is sometimes precise and can have a definitive meaning is welcome, but it is logically incompatible with an insistence on deference and judicial restraint.

If the Vavilov court had wanted to limit deference to cases of genuine interpretive uncertainty, it ought to have followed Chevron in clearly asking courts, first, to identify such cases, and then, and only then, to defer. That, of course, runs the risk of deference being relatively rare ― a risk highlighted by Justice Scalia in a lecture on “Judicial Deference to Administrative Interpretations of Law“:

One who finds more often (as I do) that the meaning of a statute is apparent from its text and from its relationship with other laws, thereby finds less often that the triggering requirement for Chevron deference exists. It is thus relatively rare that Chevron will require me to accept an interpretation which, though reasonable, I would not personally adopt. (521; emphasis in the original)

Conversely, if the Vavilov court was serious about deference-across-the-board being required as a matter of respect for legislative choice, it should have doubled down on the earlier view that statutory language inherently fails to determine legal disputes. This, in my view, would have been madness, but there would have been method in’t.

The trouble is that, as I said in my original comment on Vavilov, the majority opinion is a fudge. Collectively, the seven judges who signed it probably could not agree on what it was that they wanted, other than a compromise, and so did not want anything in particular. And so we get a judgment that, in a space of three short sentences, requires judicial review to embody “the principle of judicial restraint” while being “robust”, [13] and insists on deference while stressing that there may well be only one reasonable opinion to defer to.


Different people, and different legal cultures, will find their own ways to dance to the same tune of judicial resignation before the administrative state. Perhaps we should regard their different solutions as mere curiosities, objects of wonder but not judgment. But I don’t find this new Canadian hit, Chevron on 2, especially elegant or exciting. Not that I am a devotee of the on 1 original; but its steps at least come in a logical sequence. The on 2 version demands, as it were, that judges step forward and backward at the same time, and, with all due respect to the Canadian judiciary, I am not sure that it ― or, anyone else, for that matter ― is quite capable of such intricate footwork. Toes will be crushed, and partners disappointed if not injured, before someone realizes that the music needs, at long last, to stop.

CBC v Ferrier, 2019 ONCA 1025: Considering Consideration of the Charter

Part II of a two-part series on Doré.

 

Yesterday, I wrote about why Doré was under stress in the aftermath of Vavilov. Today, I write about a new case out of the Court of Appeal for Ontario (per Sharpe JA) that demonstrates why Vavilov means that Doré is sitting in a tense situation. While Ferrier should not be taken as the death knell for Doré—or even an indication of such—it is an indication of the tension that Vavilov arguably introduced into the world of Doré.

In Canadian Broadcasting Corp v Ferrier, the question involved “the openness of police board hearings” [1]. Under the Police Services Act, s.35(4), subject to certain exceptions, “police services board hearings are presumptively open to the public” [3]. In other words, section 35(4) sets out the test for whether a hearing should be closed.  In this case, the relevant decision-maker decided that the hearing should be closed. The CBC and others argued that the decision-maker “failed to pay adequate attention to the s.2(b) Charter right to freedom of expression by failing to require an open hearing” [4]. Specifically, the applicants argued that the so-called Dagenais/Mentuck test applied to the case: “[t]his test applies to discretionary decisions limiting freedom of the press in relation to court proceedings” [15]. The decision-maker, though, rejected the application of this test because (1) Dagenais/Mentuck apparently only applies to situations in the courtroom and (2) the relevant statute (s.35(4)) prescribed the proper test for determining whether to hold a closed hearing, and that statutory test ousted the consideration of Dagenais/Mentuck.

In addressing the standard of review, the Court was in an awkward position, because “[t]his appeal had been argued and a complete draft of these reasons had been written before the Supreme Court released its decision in [Vavilov]” [29]. Nonetheless, the Court went on to assess the standard of review under the Vavilov framework.

The main question in determining the standard of review was the proper decision under review, and the authority under which the decision was made. Sharpe JA concluded that the relevant decision was whether the Dagenais/Mentuck standard applied [32-33]. In other words, the relevant decision under review was the decision-maker’s refusal to apply the Dagenais/Mentuck test in view of the s.35(4) statutory test. To Sharpe JA, this was a decision reviewable on a correctness standard [33]. In drawing this conclusion, Sharpe JA drew a distinction (on standard of review) between cases where a Charter right was considered by a decision-maker and cases (as here) where the Charter right was expressly not considered:

[34] If the Charter rights are considered by the administrative decision maker, the standard of reasonableness will ordinarily apply.

[35] On the other hand, the refusal or failure to consider an applicable Charter right should, in my opinion, attract a correctness standard of review. As the Supreme Court explained in Dunsmuir, at para. 60, citing Toronto (City) v. C.U.P.E., Local 79, 2003 SCC 63, [2003] 3 S.C.R. 77, at para. 62: “where the question at issue is one of general law ‘that is both of central importance to the legal system as a whole and outside the adjudicator’s specialized area of expertise’ … uniform and consistent” answers are required. See also Alberta (Information and Privacy Commissioner) v. University of Calgary, 2016 SCC 53, [2016] 2 S.C.R. 555, at paras. 20-21. This is confirmed by Vavilov, at para. 17: “[T]he presumption of reasonableness review will be rebutted…where the rule of law requires that the standard of correctness be applied. This will be the case for certain categories of questions, namely constitutional questions, general questions of law of central importance to the legal system as a whole and questions related to the jurisdictional boundaries between two or more administrative bodies”.

[36]       The s. 2(b) Charter right to freedom of expression and freedom of the press relied upon by the appellants is both a matter of central importance to the legal system and a constitutional question.

In other words, Sharpe JA’s reasoning is that decisions whether to consider Charter rights at all are reviewable on a correctness standard, because such decisions are both constitutional questions and questions of central importance to the legal system, under the Vavilov framework. But once an administrator has considered Charter rights, the consideration of those rights are subject to a reasonableness standard.

Two things are notable about this distinction, taking into account pre-Vavilov precedent. First, prior to Vavilov, the decision of whether Charter rights had to be considered on the facts was not prescribed a specific standard of review by the Supreme Court, and otherwise was subject to a reasonableness standard in the Federal Courts. In Singh, for example, the Federal Court of Appeal was asked to determine whether s.110(4) of the Immigration and Refugee Protection Act [IRPA], which prescribes the conditions under which new evidence can be admitted in an appeal to the Refugee Appeal Division [RAD] exhaustively prescribed the conditions under which evidence could be admitted. An intervenor argued, for example, that “the RAD had to go beyond the requirements set out in s.110(4) and was obligated to proceed with a [Doré analysis].” [56]. However, the Court concluded (1) that s.110(4) exhaustively set out the conditions under which new evidence could be admitted, vitiating the need for a Doré analysis [62] and (2) even taking account of the fact that this argument was made, the Court ultimately concluded that the interpretation of a provision such as s.110(4) is reviewable on a standard of reasonableness [29]. This is because, among other things, the question was “not a question of law of central importance to the legal system as a whole…[23].

It is true, as Professor Daly points out on Twitter, that Sharpe JA’s approach is substantially similar to the approach adopted by the Supreme Court with respect to the duty to consult, as noted in Rio Tinto. There, a distinction was drawn between cases where the decision-maker decides whether it must consider the duty to consult (reviewable on a correctness standard) and cases where the decision-maker has consulted and it is up to the court to assess the adequacy of the consultation (reviewed on a reasonableness standard). More broadly, the distinction here—similar to the one drawn by Sharpe JA—is based on a traditional sort of test for standard of review: questions of law (existence of legal duty) are reviewable on a correctness standard; questions of mixed fact and law are reviewable on a reasonableness standard.

But the analogy to duty to consult is not entirely convincing. For one, in some cases, a duty to consult may not need to be considered by an administrative decision-maker—since the enabling statute may not mandate it (see Rio Tinto, at para 67). But Doré speaks in far more reaching terms: “Rather, administrative decisions are always required to consider fundamental values” (Doré, at para 35, emphasis in original). Following this line of thinking, Doré and its progeny have not adopted the distinction between power to consider a fundamental right and the consideration of that right, for the purposes of the standard of review (though I note Moldaver J’s reasons in Ktunaxa as approaching this bifurcated analysis). As noted above, in Singh, whether a decision-maker must address the Charter is a matter of statutory interpretation, normally reviewable on the standard of reasonableness (see also Deri, at the Federal Court, on this note).

This distinction, then, in the Charter context is not common.  Indeed, Sharpe JA seems to imply that Vavilov broadened the categories of cases in which correctness review would apply. The distinction drawn by Sharpe JA seems to give broader effect to the Dunsmuir correctness categories of “central questions” and “constitutional issues.” Take the central questions category. Following Singh, the question of whether a statute ousts the need to consider judicially-constructed tests was not a “central question” of importance to the legal system. But now, given Vavilov’s comments on the Rule of Law and the need for determinate final answers on important issues of legal interpretation (Vavilov, at para 53), it appears that there is extra grist for the mill for judges to expand the scope of the category, despite the Vavilov majority’s warnings otherwise (see Vavilov, para 61).  Moreover, on the scope of constitutional questions, and on Doré’s own holding, a distinction was not drawn between cases where a Charter argument was considered versus cases where they were not considered but should have been. The Court has never explicitly endorsed this proposition with reference to Charter rights.  Doré, instead, simply says that a decision which balances the Charter value with the statutory objective is reasonable (Doré, at para 58). Ipso facto, a decision which does not will be unreasonable, and so a decision that fails to even take account of a Charter value will be unreasonable (for an example, see Abdi, at para 30 ). But this was not a question of correctness, at least on Doré’s standard. Sharpe JA takes a different approach, relying on Vavilov.

One could make a convincing argument, then, that Vavilov changes  the pre-Vavilov state of affairs as it applies to Doré and other categories of correctness review. In other words, Ferrier eats into Doré’s domain.

That is one point, in itself. But another is that, in my view, Sharpe JA does not take the point far enough, and in failing to do so, creates a distinction that is unworkable.  In truth, the distinction between cases where Charter rights were considered and those where they were not is not a strong one on which to rest a difference in the standard of review. This is because of what Vavilov says at paras 55, 57:

Questions regarding the division of powers between Parliament and the provinces, the relationship between the legislature and the other branches of the state, the scope of Aboriginal and treaty rights under s. 35 of the Constitution Act, 1982, and other constitutional matters require a final and determinate answer from the courts. Therefore, the standard of correctness must continue to be applied in reviewing such questions: Dunsmuir, para. 58Westcoast Energy Inc. v. Canada (National Energy Board)1998 CanLII 813 (SCC), [1998] 1 S.C.R. 322.

[…]

The constitutional authority to act must have determinate, defined and consistent limits, which necessitates the application of the correctness standard.

In other words, “the constitutional authority to act”—whether Charter values are considered or not—necessitates the application of the correctness standard. More broadly, the application of the correctness standard in these circumstances “respects the unique role of the judiciary in interpreting the Constitution and ensures that courts are able to provide the last word on question for which the rule of law requires consistency and for which a final and determinate answer is necessary” (Vavilov, at para 53). In my view, this is true when a court analyzes whether a decision maker was required to consider Charter rights, and in cases where the decision-maker actually analyzed the Charter rights. In both cases, there is a substantial constitutional component to the analysis which implicates the need for the court to have the final say on the law: the court is required in both cases to assess the application and scope of constitutional rights. Even when considered in respect of facts or proportionality analysis, a court is still required to construe the scope of constitutional rights.

Some might argue with my position here. For example, as I mentioned in my previous post, Vavilov explicitly does not overturn Doré, and also does not explicitly mention questions of the “Charter” falling within the scope of its comments on “constitutional questions.” But it would be hard to distinguish between these cases. In other words, what is the compelling justification to treat Charter cases separately from all other questions of constitutional law, and going a step further, what is the justification for distinguishing cases where the Charter was considered versus where it was not? Whie one might say that the factual component changes things, in both cases, involving considering whether Charter values arise because of a relationship to a statute and cases where Charter values were considered, it is the court’s task to delineate the scope of constitutional rights. This is true in both the abstract and as applied to proportionality analysis. If this is true, the distinction, then, falls apart.

Ferrier, then, is an interesting case study in how Vavilov interacts with Doré. And at least on first blush, the interaction is tense.

After Vavilov, Doré is Under Stress

Part I of a two-part series on Doré

**This is Part I of a two part series on the interaction between Doré and Vavilov. Tomorrow, I will post a review of one of the first post-Vavilov cases, Ferrier at the ONCA. Ferrier raises issues about the standard of review on constitutional matters**

Vavilov ushhered in a new era in Canadian administrative law, particularly as it pertains to judicial review of administrative interpretations of law. That new era, as far as I can tell, is wholly inconsistent with the justifications underlying the Supreme Court’s decision in Doré in which the Court held that courts should defer to administrative decisions that engage the Charter; specifically, an administrator’s balancing of Charter values with statutory objectives. Doré is inconsistent with Vavilov in at least two ways: (1) Doré’s treatment of expertise is inconsistent with Vavilov’s treatment of the same subject; (2) Vavilov’s comments about the Rule of Law present no principled reason to distinguish between statutory constitutional questions and administrative constitutional questions (3) even if the Doré reasonableness standard is maintained, reasonableness will likely mean much more than it has in the Court’s cases subsequent to Doré . In total, what we are seeing is two cases represented by two completely different theories of administrative law. The tension is strong.

First, an admission: Vavilov hedged on Doré. This is what the Court had to say:

Although the amici questioned the approach to the standard of review set out in Doré v. Barreau du Québec, 2012 SCC 12, [2012] 1 S.C.R. 395, a reconsideration of that approach is not germane to the issues in this appeal. However, it is important to draw a distinction between cases in which it is alleged that the effect of the administrative decision being reviewed is to unjustifiably limit rights under the Canadian Charter of Rights and Freedoms (as was the case in Doré) and those in which the issue on review is whether a provision of the decision maker’s enabling statute violates the Charter (see, e.g., Nova Scotia (Workers’ Compensation Board) v. Martin, 2003 SCC 54, [2003] 2 S.C.R. 504, at para. 65). Our jurisprudence holds that an administrative decision maker’s interpretation of the latter issue should be reviewed for correctness, and that jurisprudence is not displaced by these reasons [57].

The Court was clearly right not to take the amici’s suggestion about Doré. There was no question of administrative interpretation engaging Charter rights or values on the facts of Vavilov (or the companion case of Bell/NFL), so it would be an unjustified expansion of the judicial role to deal with questions that were not before the court. But this does not mean that Doré sits easily with the new era of administrative law that Vavilov has ushered in.

Take, first, Vavilov’s comments on expertise. Vavilov concludes that expertise is not a legal reason for deference, as far as determining the standard of review [30-31]. Put differently, expertise cannot justify a presumption of deference because:

…if administrative decision makers are understood to possess specialized expertise on all questions that come before them, the concept of expertise ceases to assist a reviewing court in attempting to distinguish questions for which applying the reasonableness standard is appropriate from those for which it is not [28].

Specifically, then, expertise cannot assist a court in reviewing administrative interpretations of law, because it is not a good working assumption that decision-makers are expert on all matters that come before the court.

Now compare the tenor of these comments to what Doré had to say about expertise on constitutional matters. The Court in that case noted that a revised approach to the review of administrative decisions implicating the Charter involved “recognizing the expertise of these decision-makers” [35]. Specifically:

An administrative decision-maker exercising a discretionary power under his or her home statute, has, by virtue of expertise and specialization, particular familiarity with the competing considerations at play in weighing Charter values [48].

Putting Vavilov and Doré side by side like this illustrates the ill-fit: Doré’s underpinning concept is that of expertise, particularly the notion that administrators have expertise in legal matters as a presumptive rule. But if this is no longer assumed when it comes to selecting the standard of review in the Vavilov context, there is no reason to assume it in the constitutional context, where the case for expertise is—logically—weaker. In other words, constitutional law is not the same as routine legal matters with which administrators may have experience. Making the jump from these routine legal matters to constitutional matters was always the fatal flaw of Doré, and now that Vavilov does not even assume expertise when it comes to legal matters, there is no reason to make that same assumption on constitutional matters.

Next, consider what Vavilov had to say about the Rule of Law. The Rule of Law, according to the Supreme Court, will sometimes require “a singular, determinate and final answer” to the question before a particular court (Vavilov, at para 32). This makes sense: the Supreme Court has also said that the Rule of Law, among other things, “requires the creation and maintenance of an actual order of positive laws which preserves and embodies the more general principle of normative order” (Reference re Manitoba Language Rights, at 749). On certain questions, it would undermine this “general principle of normative order” for a court to take a “hands-off” approach to a certain decision, or to permit multiple “reasonable” interpretations of a particular issue to stand when it comes to the Constitution.

One of the categories of correctness review emphasized in Vavilov is constitutional questions, where “[t]he application of the correctness standard…respects the unique role of the judiciary in interpreting the Constitution and ensures that courts are able to provide the last word on questions for which the rule of law requires consistency and for which a final and determinate answer is necessary” (Vavilov, at para 53). Under this understanding, there is no principled reason that Vavilov’s comments about the Rule of Law and constitutional questions should necessitate a different response just because of the forum in which the constitutional argument arises. Just because, on judicial review, an applicant challenges a statute as opposed to an administrative decision should not change the task of the judiciary to be the final expositors of the Constitution. This, again, was a fatal flaw of Doré. It is, in my view, difficult for Doré to stand given Vavilov’s doubling-down on the traditional role of the courts in interpreting and applying the Constitution.

True, the concurring opinion in Vavilov (Abella and Karakatsanis JJ) was quick to point out that “[t]he majority’s approach to the rule of law, however, flows from a court-centric conception of the rule of law rooted in Dicey’s 19th century philosophy.” I’ve always been struck by the intellectual laziness of the charge of “Dicey” as a legal argument; but no matter, in the context of the Constitution, the Rule of Law is primarily the rule of courts, at least on the majority’s understanding in Vavilov. And, what’s more, the majority’s understanding is consistent with what the Supreme Court has said itself about the role of the judiciary in a constitutional democracy, in a variety of different contexts: Hunter v Southam, at 155: “The judiciary is the guardian of the constitution…”; Ell v Alberta, at para 23: “[a]ccordingly, the judiciary’s role as arbiter of disputes and guardian of the Constitution require that it be independent from all other bodies”; United States v Burns, at para 35: “…the Court is the guardian of the Constitution…”; Kourtessis v MNR, at 90: “The courts are the guardians of the Constitution and they must have the powers to forge the instruments necessary to maintain the integrity of the Constitution and to protect the rights it guarantees”; and in a judgment jointed by Abella and Karakatsanis JJ in the Nadon Reference, the Court endorsed the proposition that since the judiciary became the “guardian of the Constitution,” the Supreme Court itself became a “foundational premise of the Constitution” (at para 89). These comments can easily be taken to imply that courts, in comparison to administrators, have a unique role in interpreting the Constitution by systemic legal design, even if administrators, in the odd case, may have something of value to say about the Constitution. Particularly apt, on this score, is the comment in Ell regarding independence: courts are the only independent guardians of the Constitution.

Finally, there is something to say about reasons, even if correctness is not the applicable standard of review. Applying the reasonableness standard in the Doré context requires proportionality (see Doré , at para 56), but Doré does not explain explicitly what is required in terms of reasoning. That said, the Court has been reticent to adopt a formal reasons requirement for Doré -type decisions, consistent with the Court’s jurisprudence that reasonableness means different things in different contexts (see Catalyst, at para 18), and that the adequacy of reasons is not a standalone basis for review (see Newfoundland Nurses, at para 14), with courts being permitted to supplement reasons for decision (Newfoundland Nurses, at para 12). The classic example of this was in TWU, where the majority of the Court, relying on these authorities, concluded that simply because the Benchers were “alive” to the Charter issues, there was no issue of discretion fettering when the Law Society ordered a referendum on TWU (TWU, at para 56). In dissent in TWU, Brown and Côté JJ would have required more in the way of justification from the Law Society, especially given the Charter rights at play (TWU, at paras 295-296).

Vavilov says something different about what reasonableness requires, putting stress on Doré and its progeny. The Vavilov framework withdraws from the “supplementation” of reasons (Vavilov, at para 96), still permitting the reviewing court to look to the record, but not permitting courts to gin up its own reasons for decision. But reasons take on an expanded importance in Vavilov, specifically requiring a decision-maker to justify decision in relation to particular legal constraints on the decision-maker (Vavilov, at para 108) and in terms of the impact on the affected individual (Vavilov, at para 133), among other things. Interestingly, Vavilov does note that its reasons first methodology will be difficult where reasons are not provided (Vavilov, at para 137), explicitly citing TWU, and further notes that in the absence of reasons, courts must still apply the various constraints on the decision-maker, but that the analysis may focus more on outcome that on reasoning (Vavilov, at para 138).

Even with these comments in mind, the past precedent, TWU, and Vavilov do not stand easily together. Specifically, the most important legal constraint on any decision-maker is the Constitution. It is difficult to see how a decision-maker could fail to justify a decision under the Constitution and for a court to rule that that decision is reasonable—courts should not cooper up bad or non-existent reasoning in these cases. This is even more so given that one of the constraints on the decision-maker, the impact and importance to the affected individual, is particularly acute in situations involving constitutional rights. It might appear that in the constitutional context, more should be required if we retain a reasonableness standard on constitutional matters. Simply put, reading TWU and concluding that the decision-maker was “alive” to the Charter issues seems to be the wrong line of thinking, with Vavilov. In this sense, there are genuine signals pointing to Brown and Côté JJ’s dissent in TWU, where they would require more in term of reasoning. That said, this is an area of genuine ambiguity that I cannot resolve here, and there are also signals in Vavilov that cases like TWU are still good law.

I do not make these comments with the naïve understanding that the Court could not save Doré in light of Vavilov. Vavilov explicitly does not mention the Charter in the class of cases to which correctness should apply. And the Court’s approach could, admittedly, allow for Doré to stand. For example, courts could continue to apply the reasonableness standard to constitutional questions where expertise is demonstrated by an administrator in interpreting the Constitution. This finds some support in Vavilov, where expertise is “folded into the new starting point and, as explained below, expertise remains a relevant consideration in conducting reasonableness review” (Vavilov, at para 31). But this runs into a number of problems: first, it continues to apply a reasonableness standard even though expertise is no longer is a reason to apply that standard under Vavilov. Second, it would require courts, in each case, to measure expertise as an empirical matter on constitutional questions. While rules of thumb could be used to assist in this matter, it is unlikely to be an attractive option to the Court.

At the very least, Doré now stands at odds with Vavilov, its underlying justifications under stress because of the new administrative law foundation introduced by the Supreme Court. In my view, the two cases represent two different visions of administrative law. On one hand, Vavilov is indeed a move towards Diceyanism (and I mean this in the best way possible), in the sense that the statute is the centrepiece of the analysis when it comes to selecting the standard of review and applying it. Doré is based on more functional concerns, notably, expertise. There is a fundamental mismatch here. How long Doré lasts, only time will tell. But there is at least some reason to think that Doré is under significant tension because of Vavilov.