The Post-Vavilov Supreme Court and Administrative Law

Reason for optimism?

After the Supreme Court’s recent decisions in Abrametz and ESA (both of which are summarized and analyzed in my newsletter here and here, respectively), there is much to say.  But I just want to quickly identify one emerging trend: the centrifugal force of the principles in Vavilov in areas of administrative law not immediately in its contemplation. For reasons I hope to outline in future work, I think this trend is positive, because the sweeping and comprehensive approach in Vavilov provides a set of plausibly, administrable rules that reduces the amount of time spent on finding the standard of review, among other things. This is a good thing because the law of judicial review shouldn’t be overly complex. It is designed to be quick, on the record, and facilitative for individuals challenging government action. Labyrinthine doctrine stifles that purpose.

 For now, though, I only write to highlight a few examples of this trend I identify:

  • In Abrametz, the question of the standard of review for procedural fairness issues arising in a statutory right of appeal was at issue. For a majority of eight, Rowe J applied Vavilov’s holding on rights of appeal (the appellate, not judicial, standards apply) to the issue of delay (Abrametz, at para 27). This was despite the majority’s acknowledgement that this “proposition was stated in the context of substantive review.” This was also despite the fact that the orthodox view is that correctness applies to issues of procedural fairness (set out in cases like Khela). As I outlined in my newsletter, I think this move is justifiable, and it may raise questions about assimilating all issues of procedural fairness to Vavilov’s rules and standards. This is not an argument I can explore here, but it has been mooted in the courts (e.g. Maritime Broadcasting, at para 50). That said, one will likely find a resolute voice in favour of hard-line correctness review on all issues of procedural fairness: see Côté J’s dissent in Abrametz. And the move made in Abrametz regarding rights of appeal is a much easier hill to climb than a full-on application of Vavilov to issues of procedural fairness.

  • In ESA, the question was whether the laws like the Copyright Act that confer jurisdiction over the same questions to courts and the Copyright Board invite the correctness standard. In Rogers, a previous Supreme Court case, the Court recognized this as an appropriate circumstance to deviate from a reasonableness standard. For a seven judge majority, Rowe J again approached the problem by asking how Vavilov altered the Rogers standard (ESA, at paras 24-25). This was because—as I argue—Vavilov “simplified the law” (ESA, at para 24) and “overtook the prior jurisprudence” (Vavilov, at para 14). Analyzing the problem from the perspective of Vavilov’s principles of legislative/institutional design and the rule of law, Rowe J recognized the Rogers exception as an additional category that would attract correctness review. It is “analogous” to a statutory right of appeal (ESA, at para 32) and inconsistencies could arise between judicial and administrative interpretation, undermining the consistency and systemic clarity required by Vavilov’s idea of the rule of law (and, I argue, the Supreme Court’s other precedents). Karakatsanis J and Martin J said the majority’s conclusion “undermines Vavilov’s promise of certainty and predictability” (ESA, at para 117) because despite the fact that Vavilov “obviously considered” cases like Rogers, it did not choose to recognize it as an example of a case requiring correctness review (ESA, at paras 117, 124). In fact, if anything, it implicitly (but not “inadvertently”) overruled it (ESA, at para 125). While I do not agree with the concurrence’s worry about the majority’s decision, I note that the concurrence, too, sees Vavilov as a simplifying, cohering mechanism—it just reads Vavilov differently (in my view, much too narrowly—see Vavilov at para 70, which focuses on the application of its principles in future cases).

These Supreme Court examples only serve to buttress trends in the lower courts. In Portnov, for example, Stratas JA in the Federal Court of Appeal held that the bespoke standard of review analysis for regulations set out in Katz should be foreclosed and assimilated to Vavilov’s reasonableness standard (Portnov, at paras 24-28). And even beyond the grounds of substantive or procedural review,  Boivin JA in Ermineskin Cree Nation merged considerations of judicial discretion and cost-expediency in the question of whether a case was moot; in so doing, it tied an important consideration underlying Vavilov’s discussion of remedies on judicial review to the preliminary issue of mootness (see Ermineskin Cree Nation, at para 41; Air Canada, at para 14). What’s more, there is debate in the lower courts, and even at the Supreme Court, about the standard of review on arbitral appeals: whether the standard is reasonableness or the appellate standards, as prescribed by Vavilov (see here).

In all of these examples, we see Vavilov potentially doing a lot of conceptual work.  Not only are its principles affecting substantive review (ESA), but also preliminary objections to judicial review (mootness), and further, even in domains not necessarily within the Vavilov Court’s express contemplation (Abrametz). Nonetheless, it appears as if a number of judges on the Court are viewing Vavilov not only as a good encapsulation of accepted administrative law values (primarily, institutional design and the rule of law, but also discretion and cost-effectiveness), but a plausible set of operational rules deduced from those values.

For my own reasons, I think the overall effort in Vavilov was sound. In theory, there are some imperfections in Vavilov. I’d much prefer, for my part, a Chevron-like approach to judicial review of administrative action, which doesn’t start from a presumption of deference, but which asks courts to interpret whether the statutory language can support “more than one answer” in the first place. Nonetheless, I can live with Vavilov because it is a clear rule with clear exceptions, and we get something like Chevron  on questions of law when we apply the reasonableness standard (though this does not solve the fundamental issue: see Leonid Sirota here). The presumption of reasonableness is rooted in a plausible, though imperfect, conception of legislative delegation (again, I have my own fundamental objections to it—delegation does not necessarily equal deference). But the “wrongness” of the presumption, in my view, is mitigated by sound exceptions to its application in cases where there is a clear contrary signal “subtracting” from the specific delegation (rights of appeal, concurrent jurisdiction); where the question at issue transcends the legislative delegation (constitutional questions, for example); or perhaps even when it is unclear the delegation extends to the decision-maker at all (see here). And as I say, even on questions where reasonableness applies, we get an approach that takes into account administrative legal errors, statutory language, and which forces more extensive justification to facilitate judicial review on the reasonableness standard.

We live in an imperfect world, and in my view, this old trope extends with even more force to the law of judicial review. So at the end of the day, imperfections in the theory and doctrine do not undermine the workability that Vavilov has achieved. Because selecting the standard of review is categorical, theoretical imperfections aside, the process is much simpler. As I say above, certainty and predictability in the law of judicial review should be built-in, so that citizens can understand how to challenge government action. What I think Vavilov has done is settle the fundamentals, allowing development around the edges, while providing a sound starting point for that development. For this reason, I agree with the overall trend: courts should look to Vavilov’s principles to infuse, to a greater or lesser degree, the doctrine of judicial review where relevant.

Some Major Questions About Major Questions

In West Virginia v EPA, the Supreme Court of the United States, wielding the “major questions doctrine” found that the EPA did not have the statutory authority to adopt regulations implementing the Clean Power Plan, initially proposed by the Obama administration in 2015.  In this post, I describe why I think this decision was ultimately misguided, how the major questions doctrine might be recast, and why some of the dissent’s concerns are themselves misguided. In short, while I share concerns about agency opportunism, I do not think this judicial creation is the solution.

Under the Clean Air Act, the EPA has broad authority to establish regulatory “standards of performance” in relation to certain categories of pollutants. These include the adoption of the “best system of emission reduction,” taking into account various factors, that the EPA administrator “determines has been adequately demonstrated.” Under this provision, the EPA adopted a “generation-shifting” system rather than a plant-by-plant regulation system; in other words, it decided that the “best system of emission reduction” included generation-shifting. A generation-shifting approach, as I understand it—such as cap-and-trade, or investment in alternative sources—is a different sort of regulation because it purports to operate on a system-wide basis, rather than on an individual source basis. The EPA’s regulatory choice thus cut rather broadly to attack emissions across different sectors. There is much more detail in the opinions that I am leaving out for the sake of brevity.

The majority opinion (Roberts CJ) concluded that this was a “major question,” meaning there was “every reason to hesitate” before concluding that Congress meant to confer this authority on the EPA (20, see also Brown & Williamson, at 159-160). In order to draw this conclusion, Congress would need to include special authorization, or a so-called “clear statement.” In “major questions” cases, the majority tells us that “both separation of powers principles and a practical understanding of legislative intent” make it reluctant to find that the EPA’s stated authority here was buried in “vague” provisions like s.111(d) (19); here we see some concern about agency opportunism and attempts to expand statutory authority. The Court tells us that it has, as a matter of course, deployed the major questions doctrine in extraordinary cases where the agency asserts broad authority of economic and political significance (17). Put simply, as the majority frames it, the major questions doctrine is like a substantive canon of construction: the Court proceeds on its own assumption that Congress would not have wanted the EPA to have this authority, absent an explicit statement otherwise.

In dissent, Kagan J attacked the majority’s use of the major questions doctrine. Chiding the doctrine as a “get-out-of-text-free card” (28), Kagan J reasoned that the statute here was broad, but not vague (8). The breadth of the statute was by design. While this was so, Kagan J noted that the grant was also constrained: in concluding what the best system of emission reduction is, the EPA must consider “costs and non-air impacts” while making sure that the best system has been “adequately demonstrated” [7].

For Kagan J, the invocation of the major questions doctrine elided this delegation of power. She reasoned that the “doctrine,” such as it is, had only been deployed as part and parcel of ordinary statutory interpretation, determining the scope of delegated power conferred on an agency: “the text of a broad delegation, like any other statute, should be read in context, and with a modicum of common sense” [13]. Kagan J was concerned that the use of this special rule short-circuited the Court’s duty to actually determine what the statutory provisions meant, and whether the EPA’s regulation fit within the provision. Kagan J goes on to wax poetic about the importance of expertise and the administrative state (I’ve registered my opposition to these particular, age-old arguments before, and renew them to some degree below).

That said, I see a few problems with the majority’s recast of the major questions doctrine:

  • Despite the majority’s protestations otherwise, there has been a shift in the way the major questions doctrine works. Scholars, as I read them, tend to disagree about which cases mark the turning point, but I think it is fair to say that WV v EPA is the capstone. As I understood it, the doctrine was previously attached to Chevron deference. If a question was a “major question,” it was a reason to deny deference to the agency’s asserted interpretation; not a presumptive rule against ordinary agency action in certain, ill-defined “major” areas. But now the doctrine has taken on a life of its own; gone is Chevron. The lineage of this change does not suggest that it is justified.

 In Brown & Williamson, for example, the Court conducted a normal statutory analysis as prescribed under Chevron Step 1, and concluded that “Congress has clearly precluded the FDA from asserting jurisdiction to regulate tobacco products” (Brown & Williamson, 126, 132). It was only after this exhaustive analysis that the Court additionally concluded that, given Congress’ structured and deliberate scheme for tobacco regulation, the FDA could not likely have been given the authority to regulate tobacco products (Brown & Williamson, 159-160). This was not a standalone, substantive canon of construction; it was a tool of judicial common sense, an insight drawn from the application of the tools of interpretation, to determine that a statute precluded the agency’s view. Second, in Utility Air, the majority (Scalia J) questioned whether the agency’s construction could be reasonable under Chevron Step 2, given its “claim to extravagant statutory power” (Utility Air, at 20). Here, the “major questions” concern operates as a sort of defeasible outer limit that limits the range of reasonable options at Chevron Step 2. Even in King v Burwell, the Court used the major questions exceptions as a means to an end: to conclude that it should conduct an independent assessment of the statute in that case. In all of these cases, the major questions doctrine was operating more as an “add-in,” or “tie-breaker”: see, relatedly, fn 3 of the concurring opinion of Gorsuch J.

So, quite aside from the sotto voce overturning of these, weaker versions of the major questions doctrine, the consequences of moving from a deference rule (a rule changing the intensity of review, but not changing the legislature’s delegation process) to a delegation rule (a rule denying a power to delegate straightforwardly on an ill-defined set of questions at all)  is not small, and there are two. First, because the Court only offers a set of mushy guidelines for what constitutes a “major question” on which Congress will require an explicit statement, Congress may be left wondering how and when it must make itself “clear.” Second, on principle, Congress regrettably must—according to the Court’s standards—make itself clear. This was not required under Chevron. As Kagan J notes, the major questions doctrine—as deployed by the Court in WVA v EPA—basically shortcircuited this statutory analysis. Instead, rather than determining whether the statute supported the EPA’s reading, the Court was rather results-oriented: because the issue is big, even normal statutory authorization should not count, even though the EPA’s view was plausible (and certainly not vague—as Kagan J says, the issue here is breadth). It is plausible, as the majority suggests, that Congress did authorize this power in the text of its law. But this authorization was not explored, or otherwise, was inexplicably not enough.

  • Gorsuch J, in a concurring opinion, attempted to avoid this conclusion by stating that substantive canons of construction are accepted tools of interpretation. Indeed they are. But one should evaluate the lineage and triggers for these substantive presumptions, and at any rate, all should agree they should be used with caution and rooted in consistent doctrine and principle. Otherwise, they can be manipulable “get out of text free” cards. For this reason, many of the substantive canons are rooted in clear constitutional concerns: for example, some of the presumptions concerning federalism.

What is the constitutional basis of this form of the major questions doctrine?  One candidate might be the same pool of principles that grounds the non-delegation doctrine. But I think it is hard to justify this doctrine as some extension or analogue to the non-delegation doctrine (cf Cass Sunstein) in the mould of the Benzene Case. Undoubtedly, one can conceive of some important similarities—the Court has sometimes gestured to an asserted power as being too broad to justify Chevron deference. But, again, in all the cases, the problem was simply that Congress had foreclosed such extravagant regulation. Instead, the non-delegation doctrine, as a constitutional doctrine, is primarily concerned with the scope of delegated power. Courts seek an intelligible principle to determine whether the power is adequately guided. The idea is that, absent a guiding, legislative principle, an agency is exercising legislative power unconstitutionally, in a manner contrary to Article I.

While there are powerful historical arguments against the non-delegation doctrine that I am not equipped to evaluate, I take it as a given that it is alive and well, and that it—at least in some part—aims at preserving legislative power and control over administrative decision-making. Nonetheless, the new major questions doctrine deployed by the Court in WV v EPA is not, at least primarily, concerned with the breadth of delegated power per se. As Kagan J recounts, the delegation of power to the EPA in this case is, indeed, broad, but it is not beyond the pale. The EPA does have some meaningful constraints built-in to the statute, and such constraints likely save the delegation from any challenge on non-delegation grounds. The major questions doctrine, then, isn’t so concerned with the breadth of delegated power as much as the significance of the issue at hand. But issue significance does not necessarily equal broad delegated power; in other words, we can have broad delegated power in unimportant areas, which could plausibly raise non-delegation concerns, or we can have narrow delegations in important areas, which would raise major questions concerns. These doctrines are aiming at different things. As John Manning so eloquently said: “If the point of the nondelegation doctrine is to ensure that Congress makes important statutory policy, a strategy that requires the judiciary, in effect, to rewrite the terms of a duly enacted statute cannot be said to serve the interests of that doctrine.”

In whole, the result of this doctrinal shift is a complication of the basic task of the law of judicial review, and an arguable corruption of the legislation delegating the power at issue. The law of judicial review is designed to be an adjunct branch of statutory interpretation, to determine the scope of powers granted to an administrator. Whether Chevron has one or two steps, this is the core of the thing. With this new major questions doctrine, the law of judicial review is somewhat different. Less important is the text of the laws delegating power. More important is whether the court thinks the issue over which the power is delegated is “important” or “big” enough. This abstracts away from the core question on judicial review: does the agency have this asserted power to conduct this action, no matter how big the problem  may be?

II.

With these concerns in mind, I think it might be useful to consider how Canada deals with questions of this sort, because I think the doctrine aims at some of the same concerns as the major questions doctrine while avoiding some of the potential pitfalls, as I see them.

 In Canada, our going-in presumption is reasonableness review (for an apt description of why Canada gets this backwards in relation to the United States, see Leonid Sirota). This presumption can be rebutted, in which case the court reserves to itself the right to pronounce on the legal issue without any deference. One of the circumstances in which the presumption can be rebutted is in cases involving “general questions of central importance to the legal system as a whole” (see Vavilov, at para 58 et seq). This category has much in common with the gist behind the “major questions” category, but there are important differences. First, the justification for the central questions category in Canada is not concern about delegated power; rather, it is designed to protect the role of the judiciary as the most important external check on administrative power. Issues that are considered centrally important must be answered consistently by courts, because they engage the rule of law, over which the judiciary is the primary guardian. Notably, however, in defining a central question, the courts are careful not to unduly eat away at the legislature’s right to delegate power. As the Supreme Court says, “the mere fact that a dispute is ‘of wider public concern’ is not sufficient for a question to fall into this category—nor is the fact that the question, when framed in a general or abstract sense, touches on an important issue” (Vavilov, at para 61).

And so the questions recognized by the court concern the stability of the legal system, issues that transcend an administrator’s particular statutory grant of authority. These issues have typically concerned constitutional or quasi-constitutional issues. For example, in University of Calgary, the question was whether a decision of an administrator that a statute permits solicitor-client privilege to be set aside. The Court refused to defer in this case, because solicitor client privilege has a constitutional dimension, and the “question of what statutory language is sufficient to authorize administrative tribunals to infringe solicitor-client privilege is a question that has potentially wide implications on other statutes” (University of Calgary, at para 20).

Why is this doctrine preferable, at least on these points, to the major questions doctrine, as deployed in WV v EPA? For one, the Canadian version expressly forecloses the possibility that an issue’s importance factors into the analysis. Many issues can be characterized as important; this is the challenge of the law of judicial review, to channel and sometimes restrict administrative power over these important areas of public life. Second, the benefit of a rule restricting deference is that it can be justified with reference to fundamental tenets of the law of judicial review. There are good rule of law reasons, pertaining to the judicial duty to pronounce the law, to carve away deference in cases where an agency attempts to exert power in a way that may transcend its own statute, or engage quasi-constitutional norms. Legislatures should not be impliedly granted the power to, by delegation, carve away the core powers of judicial review. And so,  uniformity of the administrative justice system requires not percolation, but correct answers. This approach also answers for the problem of agency opportunism: an agency that seeks to manipulate its own powers in order to affect these broader areas of legal importance should be met with a resolute judiciary. Nonetheless, the category is pared down not to every conceivable important issue, but to core concerns of the judiciary that implicate the rule of law. With these more narrow justifications, I would venture that the Canadian law of judicial review—only on this score—may be more stable than the American.

III.

In the meantime, I do not expect nor suggest that American courts should look to Canada for guidance. Even still, I do not think that the US situation is as dire as the dissent and some commentators suggest. As Kristin Hickman notes, Congress can rectify the ruling in WV v EPA tomorrow (even if it should not have to): it can simply legislate a clear statement. To be clear, this is an additional, probably unjustified hurdle. But the ball is in Congress’ court.  And, as I have argued before, non-delegation limits on administration will not hobble administration, and may actually incentivize better deliberation and guidance.  Further, administrative government is vast, and few regulations are challenged such that this case will make a difference on a wider scale. These are the practical realities that limit this new major questions doctrine, but I admit there may be more I am simply missing or misunderstanding

Nonetheless, as far as it goes, the new major questions doctrine that matured in WV v EPA is unwelcome from my perspective. Unlike the non-delegation doctrine, it cannot easily be traced to the same constitutional concerns, and it is isn’t immediately clear to me that “clear statements” will really provoke much deliberation. The doctrine appears to be a wholly judicial creation; it complicates the law of judicial review, puts a hurdle up for Congress, and finds no real animating purpose, beyond nods to concerns already covered by the existing non-delegation doctrine. I stress, as I did above, that I too worry about “bureaucratic domination,” drift, etc. But these are concerns that are best kept in check by a legislature and by consistent and principled application of the law of judicial review. Creating an amorphous thumb on the scale, in this manner, is an ill-fit.

Immigration and Refugee Decision-Making: The Vavilov Effect?

It has been a while since I’ve blogged. The last few months have been—in a word—chaotic. I’m hoping to blog more regularly going forward now that some of these things have settled

One of the areas where administrative law really comes to life is in immigration decision-making, particularly front-line decision-making like visa decisions or humanitarian and compassionate decisions [H&C]. This is where the pressures, incentives, and moral worldview of “street-level bureaucrats” in particular contexts can tell us about how decisions affecting all-too-real rights and interests are made. The area, though, presents all sorts of challenges for those studying the law of judicial review.

First, immigration visa decision-making is also just one particular iteration of a broader reality: the inexplicable diversity of administrative decision-making. That diversity leaves monist accounts of the administrative state wanting. Expertise—advanced by the Progressive school as a core reason for delegation and deference—presents a different empirical reality in these contexts. In other words, this is not the labour board or the human rights tribunal where we might have more confidence in the “expert” nature of the decision-maker. In this context, not only is “expertise” not to be assumed, but what it means on the frontlines escapes easy definition.

Second, emerging democratic theories view the administrative state either as a place to facilitate and channel democratic deliberation or a place to encourage contestation (agonism). These theories are deeply insightful and may have resonance in other areas. But in some of these immigration and refugee cases, it is hard to say that there is anything substantively democratic happening. The only democratic argument is entirely formal: the delegation of power to officials to make decisions. This delegation of power must be respected, but the chances for contestation or facilitation seem far off.

Other features of front-line immigration visa decision-making present problems from the perspective of the law of judicial review. Notwithstanding what I say below, it was typically the case that visa decisions did not—and still, do not—require extensive reasons: Persaud v Canada (Citizenship and Immigration), 2021 FC 1252 at para 8. And in theory, this remains true post-Vavilov. What’s more, there was, and remains, a presumption that decision-makers considered all the evidence before her: Cepeda-Gutierrez v Canada, 1998 CanLII 8667. 

The combination of these rules, to my mind, creates an important tradeoff. On one hand, given the backlogs in this area of administrative decision-making, we may think that officers should not spend time writing extensive reasons. On the other hand, a paucity of reasons or an adequate record that “immunizes” decisions from effective review presents problems from the perspective of legality, but more directly, to the individuals who wish to seek judicial relief: see Canada (Citizenship and Immigration), v Canadian Council for Refugees, 2021 FCA 72 at para 102.

There should be some balance struck here. Post-Vavilov, courts in some cases are beginning to strike this balance. They have done so in favour of more substantive reasoning that addresses the legal and factual stakes to the party affected by a decision. In other words, in these cases, the courts are not abiding boilerplate and rote recitation of the facts. Nonetheless, they are not expecting long, involved reasons in every case, and they need not be perfect: the reasons can be short, but should be directed to the actual stakes facing the individual. In my view, this decisively moves the balance towards the ideal of legality, understood in this case as enhancing the role of the courts to ensure compliance with administrative law.

Here are some examples of what I am describing:

  1.  Singh v Canada (Citizenship and Immigration), 2022 FC 692

Here, Justice Diner describes well the post-Vavilov position on reasons:

[22] Visa officers are certainly entitled to deference, but only where their findings have at least a modicum of justification. That was entirely absent here. In the age of Vavilov, the Court cannot defer to reasoning missing from the Decision, or fill in that reasoning for administrative decision-maker. Lacking justification, the matter will be returned for redetermination

2. Rijhwani v Canada (Citizenship and Immigration), 2022 FC 549

This was a denial of a permanent residence application where the applicant plead H&C grounds. The applicant specifically pointed to establishment and hardship as supporting her application. The Officer did not address these factors in detail. The Court says, at para 17: “It is particularly important that when there are few factors raised—in this case only hardship and establishment—that the Officer addresses the rationale clearly for each.”

This did not occur here. Noting, at para 10,  that “brevity cannot excuse inadequacy” the Court takes issue with the “two significant errors…in under a page of reasons” that characterized this decision.

3. Gill v Canada (Citizenship and Immigration), 2021 FC 1441

Gill was found inadmissible to Canada for five years by a visa officer because of misrepresentation; he failed to disclose an unsuccessful tourist visa application to the United States. Gill advanced the argument that his “misrepresentation” was actually an innocent mistake. He argued that the officer did not reasonably explain why he rejected the “innocent mistake” argument.

Specifically, the officer in this case apparently took—word-for-word—reasons that were given by a separate officer in another case that was reviewed in the Federal Court. Speaking of the Cepeda-Gutierrez presumption, the Court said, at para 34:

I note, however, that the use of identical template language to express not just the relevant legal test or framework, but the reasoning applicable to an applicant’s particular case undermines to at least some degree the presumption that the officer has considered and decided each individual case on its merits.

The Court did note, however, that templates can be useful tools in high volume-decision-making [33].

I do not present these cases to make an empirical claim about what any number of courts are doing post-Vavilov. This is impossible to do without closer study. But I can say that there are many more of these cases, and I recommend you consult my weekly newsletter if you are interested in reading more. In the meantime, I think we can draw some conclusions from these cases:

  1. There is something to be said for a signal sent by a judicial review court to administrators about what they should expect. Prior to Vavilov, decision-makers may have expected strong presumptions of deference and courts claiming that inadequate reasons did not provide a standalone basis for review. Now, decision-makers may expect a closer look if their decisions are reviewed, particularly in this front-line context. One hopes that this incentivizes structural solutions within administrative bodies. This should not be hard to expect from Immigration, Refugees, and Citizenship Canada, which houses Canada’s largest administrative decision-maker.
  2. No one should take this to mean that reasons need to be extensive in every case. But it should be taken to mean that boilerplate is presumptively problematic. This is because boilerplate, by its nature, does not respond to the individual stakes raised by many of the decisions in the immigration realm. This is, in part, the thinking behind the Vavilovian constraints. If the constraints bind differently in different cases—if Vavilov is truly contextual—then boilerplate is a non-starter because it will generally fail to account for the context of various decisions.
  3. Nor is this emerging line of cases overly onerous for administrative decision-makers or front-line officers. Again, the reasons need not be perfect, need not look like a judicial decision, and need not be extensive. But they must address the actual legal and factual issues at play. If a decision-maker cannot do this, then one should wonder why they were delegated power in the first place.

At any rate, this is an area that I hope receives more attention going forward.

Same Pig, Different Lipstick: Bill C-11

Last year, I wrote about Bill C-10, which was concerned with “compelling companies like Netflix Inc and TikTok Inc to finance and promote Canadian content.” The Bill was controversial, not least because the law could be read to target content produced on user-driven sites (TikTok, say) targeting individual content creators rather than the tech giants and subjecting them to discoverability requirements and penalties. One of the biggest concerns was free expression. This law could be read to grant Canada’s telecom and broadcast regulator (the CRTC) power to regulate the content of individual expressions, something that—to many of us—presented constitutional and regulatory concerns. As Professor Michael Geist of the University of Ottawa stated upon the tabling of the bill, it “hands massive new powers to Canada’s telecom and broadcast regulator (the CRTC) to regulate online streaming services, opening the door to mandated Cancon payments, discoverability requirements, and confidential information disclosures, all backed by new fining powers.” 

Bill C-10 died because of the election, and some of us thought that would be the end of this. Not so. Yesterday the Trudeau Government re-introduced the same pig with different lipstick: Bill C-11. Professor Geist has led the charge on this and I would direct you to his site for deep analysis of the Bill, but for now, it’s enough to say that this Bill is generally not an improvement on its predecessor, at least from the perspective of the power it vests in the CRTC. Its central problem is hinging the entire controversy of the Bill on a clause which allows the CRTC to decide when and to whom the Act applies, subject to some exceptions. This should be, if not constitutionally problematic, politically so: this is the power to expand the scope of the law to a large class of individual users, allowing the Government to evade responsibility for this controversial choice in Parliament. In other words, the Government still has power to regulate user generated content and subject that content to discoverability regulations and users to potential penalties. It has this power despite the Bill representing that it does not.

Let’s take a look at the Backgrounder for the Bill. The Government says that this Bill solves two problems with Bill C-10. First, “it captures commercial programs regardless of how they are distributed, including on social media services.” Second, “the proposed bill is also clear that the regulator does not have the power to regulate Canadians’ everyday use of social media, including when they post amateur content to these services.” It seems, then, that the proposed bill does not apply to Canadian users or individual creators. And the opening part of the actual text of the Bill sounds promising. It says that it must be construed and applied in a manner that is consistent with “(a) the freedom of expression and journalistic, creative and programming independence enjoyed by broadcasting undertakings.” Section 4.1 (1) of the Bill sounds even better: “This Act does not apply in respect of a program that is uploaded to an online undertaking that provides a social media service by a user of the service for transmission over the Internet and reception by other users of the service.” This seems to deal with the problem so many of us had with Bill C-10 when it purported to extend its scope to the average TikTok user.

This sounds like a real improvement. But the promise fades when we consider the CRTC’s new regulation-making power. A regulation is a form of law—the power to make regulations is given to an agency by the elected legislature. This isn’t itself inherently problematic, and of course regulation-making is widespread today. But this goes further. Section 4.1(2) of the Bill basically “takes back” s.4.1(1), when it gives the CRTC power to make regulations governing “programs” despite the seeming exclusion of user content. This is something approaching–if it isn’t already–a Henry VIII clause, which allows an agency to amend a primary law (h/t Leonid Sirota for raising this point). If not constitutionally problematic, it is politically so. It allows the Government to evade responsibility for the potentially vast scope of this law.

This is the controversial clause. It is cabined by a few factors, namely s.4.2 (2) (a) which directs the CRTC to consider “the extent to which a program, uploaded to an online undertaking that provides a social media service, directly or indirectly generates revenues” as it makes regulations. As Professor Geist notes, the target here appears to be YouTube music. But there are many other types of user-generated content that could conceivably fall under the scope of the law, including user generated TikTok videos or podcasts that indirectly generate revenue and have other features that fall within the scope of the regulation-making power.

The end result, as Professor Geist says, is that this technical change “would likely capture millions of TikTok and YouTube videos.” In his post on the Bill, he summarizes the wide berth of power granted to the CRTC in Bill C-11:

Views on the scope of this regulatory approach may vary, but it is undeniable that: (1) regulating content uploaded to social media services through the discoverability requirement is still very much alive for some user generated content; (2) the regulations extend far beyond just music on Youtube; (3) some of the safeguards in Bill C-10 have been removed; and (4) the CRTC is left more powerful than ever with respect to Internet regulation.

Taking into account alternative views on the scope of the Bill, I agree. The Bill basically downloads the real decision-making a level down. Rather than the Government taking responsibility for regulation user content in this fashion, it will grant it to the “independent” CRTC. If there is controversy about a future regulation, the Government can shift responsibility to the CRTC. The regulation-making just reinforces this, granting a power to the CRTC to expand the scope of the law and to make the decisions Parliament should be making in plain view.

Others will differ. They could say that I am discounting the CRTC’s own democratic process. Or, one might say that the statute cabins the regulation-making power, and that the income-generation factor is one, non-exhaustive factor. Maybe they’d be right. But I think I could grant all of this and still maintain that the Bill purports to grant significant power to the CRTC to apply the law to users, something the Backgrounder suggests it does not. This disparity concerns me.

It is important here to address another possible response. Much is made in administrative law about the need to empower regulatory experts to make decisions in the public interest. So far as this goes, the device of delegation could be useful. But it is not always and everywhere so, and there are differences in kind. A delegation to the CRTC here may be justifiable, but the Government should take responsibility for the choice to regulate user content. Presumably, this should be something that—if it needs to be addressed—should be addressed in the primary law, rather than by the CRTC in its own wide, relatively unconstrained discretion. In other words, if Youtube music is the problem, the law should be appropriately tailored.  And the use of something like a Henry VIII clause is ill-advised, to say the least.

The basic problem here might be more fundamental. I am candidly not sure what the need for this Bill is, particularly the targeting of user content. It seems the regulatory goal here may be to subject the Act’s requirements to users who generate a certain income, for example, and among other things. If that is the regulatory goal, why is the CRTC regulatory mechanism desirable here? If the Government wants to make this policy choice, why can’t it do so in the plain view?  Perhaps I simply do not understand the CanCon-motivated reason why this particular power is justified.  I’m open to someone explaining to me what I might be misunderstanding here—perhaps something specific to this regulatory context.

Nonetheless, I think there are real democratic tradeoffs to the use of this sort of regulation-making power, and more specifically the deflection of responsibility to the CRTC. This is a controversial application of a regulatory law—with penalties—to a potential huge class of users. Not only does the Government purport not to do this, but it does it here with a delegation to the CRTC. If later challenged, the Government can simply defer to the CRTC.  I do not see this legal device—and this Bill—as any better than Bill C-10.

A Cheer for Administrative Law

Administrative law can only do so much to avert injustice―but what it can do still matters

I’d like to come back, however belatedly (sorry!) to an interesting post by Paul Daly at Administrative Law Matters. Professor Daly uses the example of Novak Djokovic’s ultimately unsuccessful challenge to the Australian government’s cancellation of his visa to illustrate “the value of administrative law”. He writes:

Immigration has traditionally been a prerogative of the executive, an island of unregulated discretion in the legal system. But over the last half century or so, courts around the common law world have landed on this island, wielding principles of procedural fairness and substantive reasonableness. Several generations ago, it would have been unthinkable that a Minister would give any reasons — still less 10 pages of reasons! — to support a decision to cancel a visa. Yet because the courts now stand ready to scrutinize executive action, ministers can no longer rely on authority alone to make decisions. They must engage in the reasoned exercise of public power.

Professor Daly acknowledges that “[a]dministrative law is no panacea. Hardly any immigrant has Djokovic’s resources and will receive the Cadillac justice he has been receiving.” That’s true of course. Still, he concludes that Mr Djokovic’s case “is an important reminder of the value of administrative law in pushing ministers and others to justify their exercises of public power in reasoned terms”. This also is true. And one might even add that, in law as elsewhere, those who cannot afford the Cadillac will often benefit from the ability and willingness of others to shell out for one.

Yet despite this I think that Mr Djokovic’s case shows at least as much that administrative law is, at best, only a partial remedy to injustice. Granting the point that it can force officials to “engage in the reasoned exercise of public power” (which is often though not always true), it does comparatively little to ensure that the power is exercised justly, and nothing at all to ensure that its existence is just. The latter of course is not administrative law’s role. But it’s a point that we should not lose sight of if we choose to celebrate administrative law. In a just world, there would be a great deal less administrative law than we need in ours.

In our world, it is indeed an achievement that immigration decisions have to be reasoned and justified. After all, the founding father of Canadian administrative law scholarship evidently lamented the fact that, although the government “regards immigration as a privilege, not as a right, and wants to avoid having to disclose to a court its sources of information about the political colour of immigrants”, courts lack the good sense to see the point: “On the other side of the ideological fence, a court , with the sweating immigrant before it, sometimes sets aside a deportation order on very flimsy grounds, for instance, that it was made on a Sunday”. (John Willis, “Administrative Law in Canada” (1961) 39:2 Can B Rev 251 at 258) It’s good that we’ve moved that ideological fence some way towards decency.

But let’s not kid ourselves. We haven’t moved it very far. As Maria O’Sullivan explains in The Conversation, the reasons that ostensibly motivated the cancellation of Mr Djokovic’s visa were that his ― presumed ― opposition to vaccination against the present plague might encourage similar opposition among Australians and might undermine “social order”. Professor O’Sullivan points out that ministerial explanations were questionable on their own terms. But she also notes that, perhaps more importantly for the future, the precedent set in Mr Djokovic’s case means that people’s ability to come to Australia might be taken away on account of their actual or even perceived views being a hypothetical source of possible trouble in the opinion of a minister. What starts with an arrogant fool of a tennis player won’t stop there. Yet substantive Australian immigration law seems to allow for precisely this result, and administrative law offers no redress.

Redress will come, not any further development of administrative law, but from substantive law being such to prevent this sort of injustice. In this regard, it is telling that Professor Daly sets his reference point to 50 or 60 years ago, when immigration restrictions ― and the government’s willingness to treat immigration as a privilege to be granted or withheld on a political whim ― had become generally accepted. But let’s not stop 50 years ago; let’s go back another century. In 1872, English-speaking countries simply did not restrict immigration, though health measures and quarantines did exist. (Hence let me note: I’d have very little sympathy for Mr Djokovic if he had been barred from Australia due to not being vaccinated. But that’s very much not what has happened.) In North America, immigration controls were the product, first, of anti-Asian racism in the late 19th century, and then of more generalized xenophobia in the first decades of the 20th. On the other side of the pond, as David Cannadine writes in The Victorious Century, the closing of the UK’s borders at the turn of the 20th century was the result of bigotry against the Irish and, especially, of anti-Semitism. Australia too implemented and long held to an overtly racist immigration policy.

Of course, contemporary immigration law does not discriminate as overtly. But the idea that movement across borders is something that can be regulated in the first place comes from that evil and unjust source. And it still means that people can be stopped from doing the same (often stupid) things that we are allowed, even though they are in all particulars bar their failure to have been born in the right place or to the right parents the same as us, for no reason other than that failure. The old-school racism may be gone, but the xenophobia inherent in the idea of immigration restrictions remains. And it is not administrative law that will purge it, but the realisation that the closing of the borders 120, 140 years ago was an injustice, and that it must be ended.

Hence I will only give one cheer for administrative law. Not two, for administrative law is not meant to reform repressive substantive laws, and certainly not three, for it is powerless to mend injustice raised up to the rank of political philosophy. The trouble with cheering too loudly for administrative law is that this risks making us forget these deeper injustices; we might be content with bringing order and reason to what remains, at bottom, a logic of repression.

But my cheer for administrative law ― at least, for robust administrative law, which truly holds the administrative state to its legal and constitutional duties, rather than for the all-too-often diluted version that many administrative lawyers prefer ― will be a loud one. As E.P. Thompson famously said,

We ought to expose the shams and inequities which may be concealed beneath this law. But the rule of law itself, the imposing of effective inhibitions upon power and the defence of the citizen from power’s all-intrusive claims, seems to me to be an unqualified human good. To deny or belittle this good is, in this dangerous century when the resources and pretentions of power continue to enlarge, a desperate error of intellectual abstraction. More than this, it is a self-fulfilling error, which encourages us to give up the struggle against bad laws and class-bound procedures, and to disarm ourselves before power.

Administrative law is an essential component of the Rule of Law, and so of the unqualified human good that Thompson had the wisdom to discern amid what he saw amid great substantive injustice. Hooray for it.

Boilerplate in Decision-Making

Administrative boilerplate is probably legion in government, but of course, this is an empirical question. Nonetheless, I have read enough cases to know that individuals at the foot of administrative power—many times in front-line decision-making— are at least sometimes faced with deciphering reasons that purport to have “considered all the factors.”  Confronted, as well, with a strong presumption that decision-makers considered all of the evidence in the first place (Cepeda-Gutierrez), it is theoretically hard for applicants to move beyond boilerplate.

Besides internal administrative mechanisms that could—but may not—discourage this sort of behaviour, judicial review doctrine in Canada is starting to take notice of it. Here are a few recent cases:

Gill v Canada (Citizenship and Immigration), 2021 FC 1441

In this case, a visa officer in New Delhi used almost identical language to reject Gill’s application as another visa officer used in another denial out of New Delhi. The Court said [34]: “I note, however, that the use of identical template language to express not just the relevant legal test or framework, but the reasoning applicable to an applicant’s particular case undermines to at least some degree the presumption that the officer has considered and decided each individual case on its merits.”

Alexion Pharmaceuticals Inc v Canada (Attorney General), 2021 FCA 157

In this case, the Federal Court of Appeal chastised the Patented Medicine Prices Review Board for, among other things, “conclusory” analysis that purported to consider all the evidence [43]. This was important for the Court: “At best, on this point the Board obfuscated, making it impossible for a reviewing court to know whether the Board has helped itself to a power it does not lawfully have. By obfuscating, the Board has effectively put itself beyond review on this point, asking the Court to sign a blank cheque in its favour. But this Court does not sign blank cheques. Administrators cannot put themselves in a position where they are not accountable.”

Publicover v Canada (Attorney General), 2021 FC 1460

In this case, the Minister of Fisheries and Oceans denied a request for a  lobster licence transfer. In her reasons, the Minister stated that she had considered “all the relevant circumstances” [16]. The Court was troubled by this boilerplate, because it did not show that the Minister connected her analysis to the actual law and policy governing the decision [62, 66].

These cases represent a decisive shift from pre-Vavilov caselaw. Gone is Newfoundland Nurses, which permitted courts to take these boilerplate statements and “supplement” them: Nfld Nurses, at para 12. Underlying this doctrinal innovation was an unqualified presumption about administrative decision-making: “To me, it represents a respectful appreciation that a wide range of specialized decision-makers routinely render decisions in their respective spheres of expertise, using concepts and language often unique to their areas and rendering decisions that are often counter-intuitive to a generalist” [13].

In the context of boilerplate, Nfld Nurses makes little sense. This is because boilerplate reasons do not do anything to show expertise or the use of specialized concepts or language. It is merely a “say-so” of the decision-maker. Even on the Dunsmuir standard, it was always hard to say–with a straight face– that this sort of reasoning is “justified, transparent, and intelligible.”

Second, Vavilov’s renewed focus on justification and a “reasons-first” approach will be, I think, a boon for those challenging front-line decision-making. There are necessary caveats: reasons are not always required, and in many administrative contexts (such as high-volume study permit decision-making), “extensive reasons are not required” (see Niyongabo v Canada (Citizenship and Immigration), 2021 FC 1238 at para 12). But even in these areas, courts could be more willing to subject front-line decision-making to a slightly higher bar in terms of reasoning.

Third, I think this turn of events marks a tension between the Cepeda-Gutierrez presumption of consideration and the culture of justification endorsed in Vavilov. This tension was pointed out, as I noted above, in Gill. The presumption of consideration makes sense from an efficiency standpoint: after all, legislatures delegate to decision-makers for a reason, and when they do, courts should generally not go on a line-by-line treasure hunt for error. But at the same time, these efficiency concerns should take a decidedly second place: as noted in Alexion, judicial review becomes difficult when there is only boilerplate shedding light on an ultimate decision; this is to say nothing, of course, of the dignitarian reasons why reasoned decision-making is desirable (see, for a recent analysis of these issues, Janina Boughey).

This is all for the best. Boilerplate may work well in a “top-down” culture of decision-making in which those subject to administrative power and courts are in the thrall of purported administrative expertise. No need, on this account, for a decision-maker to show their work; the “just trust us” ethic is what governs. But Vavilov has arguably changed things: gone is the presumption of expertise, and gone should also be the presumptions about reasoning. If expertise exists, it can and should be demonstrated through persuasive and responsive reasons that allow a court to determine the legal basis of a decision.

Fizzy Drink or Fuzzy Thinking?

Questionable arguments in Cass Sunstein and Adrian Vermeule’s critique of anti-administrativism

I have finally started reading Cass Sunstein and Adrian Vermeule’s Law and Leviathan: Redeeming the Administrative State. As it says on the tin, the (very short) book is a defence of the administrative state, and of American administrative law, against criticism by those who ― like me ― would to tear it all, or at least much of it, down. Ostensibly, the book is offered as something of an olive branch, an argument for why those who suspect that the administrative state is inimical to the Rule of Law are mistaken about this, and can, if not embrace powerful government agencies vested with vast discretionary powers, then at least make peace with their existence.

But it gets off to a questionable start in the first chapter, which describes ― and pokes fun at ― anti-administrativist thinking, which Professors Sunstein and Vermeule brand “the New Coke”, ostensibly in reference to Chief Justice Sir Edward Coke, whom anti-administrativists like Philip Hamburger regard as a hero of opposition to executive-branch tyranny but presumably also to one of the biggest flops in the history of marketing. I don’t mind the jab ― it is amusing, although of course the Chief Justice’s name doesn’t sound like that Coke. I do mind that the argument is less forthright than it ought to be.

Professors Sunstein and Vermeule make two main points in their first chapter. One, which is less interesting both to them and surely to most of their non-American readers, is that there is no particularly strong reason to think that the US Constitution’s original meaning outlaws the modern administrative state. The other, in which they are more invested and which will resonate abroad (indeed they assert, in the introduction, that their argument is “promising … for nations all over the world” (18)), is that the administrative state is essential for government to do its rightful work, and that its critics are mistaken to only focus on its alleged dangers for democracy and liberty. This is what interests me here.

One argument I find objectionable has to with the relationship between the administrative state, liberty, and markets, and the relationship of the law, especially the common law, with all three. While anti-administrativists see the administrative state as a threat to be neutralized,

[f]or the theorists and architects of the modern administrative state, private power, exercised through delegation of legal powers and entitlements by the common law and by market ordering, was itself a threat to individual liberty. Hence vigorous government, checking the abuse of corporate and other private power, was deemed just as indispensable to liberty as were constraints on executive abuse. Consider, for example, the question whether the Social Security Administration, the National Labour Relations Board, the Securities and Exchange Commission, and the Federal Trade Commission are threats to freedom or indispensable to it ― questions on which reasonable people differ. (30)

There are several problems with this. First, the claim that private liberty is just something “delegated by the common law” is, at best, taking sides in a contentious debate. The common law itself did not see things that way. A person is free to do that which the law does not prohibit; he or she does not require the law’s permission.

Second, I think it’s quite fair to say that “vigorous government checking the abuse of … private power” is important. Recall Dicey’s example of Voltaire being “lured off from the table of a Duke, and was thrashed by lackeys in the presence of their noble master” and “unable to obtain either legal or honourable redress”. But to say so is not to answer the questions of what forms of “private power” can legitimately be checked by the state, and how they should be checked. Professors Sunstein and Vermeule want us to assume that refusal to deal is the same thing as a private violence in this regard, and that an administrative agency making law and adjudicating claims that the law it made has been infringed is no different from the police and independent courts enforcing the criminal law. These things don’t follow.

And third, the question Professors Sunstein and Vermeule pose is misleading. Reasonable people really should not differ on whether administrative agencies that can create rules backed by the threat of penal sanctions are a threat to liberty. Of course they are! What reasonable people can differ about is whether, all things considered, the threat is offset by, on the one hand, the good these agencies might do and, on the other, the mechanisms that might be devised for controlling and minimizing it. I think that it’s fair for them to argue that the administrative state does good things and that its critics have an unwisely single-minded worldview (whether or not these arguments ultimately succeed is, of course, a different question). But to deny that the administrative state threatens liberty is to peddle a similarly one-sided set of beliefs.

Professors Sunstein and Vermeule go on to give an example of how private law and private power threaten liberty, so that the administrative state is no more coercive than private ordering which it displaces:

If some people have a lot and other people have only a little, it is … not because of purely voluntary achievements and failures, important as those are. It is also because of what the law chose to recognize, protect, or reward. A homeless person, for example, is deprived of access to shelter by virtue of the law of property, which is emphatically coercive. In these circumstances, the creation of modern agencies … did not impose law or coercion where unregulated freedom previously flourished. They substituted one regulatory system for another. (31)

This, again, is quite misleading, and indeed the example comes close to doing the opposite of what Professors Sunstein and Vermeule intend ― it shows the dangers of the administrative state rather than its benefits. A homeless person is not deprived of shelter by “the law of property”, but by refusals to deal on the part of prospective landlords ― and possibly, at one remove, by prospective employers.

I’ll explain why the difference matters presently, but first, it’s important to see that the “law of property” would just as happily assure a person of a home as deny them one. Indeed, when we consider how attempts to interfere with the law of property have fared, we can see that, if anything, it would much rather provide shelter to everyone, as it were. Attempts to abolish private property in land and housing in the Soviet Union did not eliminate homelessness ― but they did result in a dire shortage of housing, such that multiple families were forced to share “communal apartments” with a handful of others if they were lucky, and with dozens if they were not. (My mother was born in such an “apartment” which her parents shared with seven other families.) Less dramatically and closer to us, administrative interference with property rights by means of zoning and building codes raises the cost of housing and prevents enough of it from being built ― which, of course, helps make people homeless in the first place.

In a competitive marketplace, refusal to deal by a prospective landlord or employer will seldom condemn a person to homelessness. Because landlords and employers compete for tenants and employees as much as the latter compete for apartments and jobs, some will moderate their demands to the point when even people who are not well off and/or have limited skills will find something for them. To be sure, some people will still need help ― temporarily in some cases, permanently in others. But this help can take the form of cash transfers, rather than regulation. But once regulation, often enacted by the administrative state, starts restricting the supply of housing or raising the cost of workers beyond what they can produce, refusals to deal by the artificially depressed number of landlords and employers risk becoming much more dramatic. In short, Professors Sunstein and Vermeule, like many well-intentioned pro-administrativists before them, are presenting as solutions mechanisms that often serve to aggravate problems they purport to solve.

This brings me to the last issue I would like to address. Professors Sunstein and Vermeule quote at length a wonderful passage from “The Federalist No. 41“, by James Madison ― a hero for many originalists and supporters of limited government whom they are eager to enlist as an ally to their cause:

It cannot have escaped those who have attended with candor to the arguments employed against the extensive powers of the government, that the authors of them have very little considered how far these powers were necessary means of attaining a necessary end. They have chosen rather to dwell on the inconveniences which must be unavoidably blended with all political advantages; and on the possible abuses which must be incident to every power or trust, of which a beneficial use can be made. … [C]ool and candid people will at once reflect, that the purest of human blessings must have a portion of alloy in them; that the choice must always be made, if not of the lesser evil, at least of the GREATER, not the PERFECT, good; and that in every political institution, a power to advance the public happiness involves a discretion which may be misapplied and abused. They will see, therefore, that in all cases where power is to be conferred, the point first to be decided is, whether such a power be necessary to the public good; as the next will be, in case of an affirmative decision, to guard as effectually as possible against a perversion of the power to the public detriment.

Professors Sunstein and Vermeule rely on Madison in support of their rejection of “a fallacious mode of reasoning that afflicts the New Coke critics of executive power” ― namely,

selective attention … to the risks of … government action, as opposed to inaction; to the risks arising from the functions of government, as opposed to dysfunctional governments … ; to the risks generated by new powers, as opposed to t he risks arising from old powers that the new powers could be used to counter. (34-35)

This is not altogether unfair: it would indeed be a mistake to only assess government institutions by the potential for abuse of their powers, without asking what good they might be able to do.

But Madison does not stop at this. His affirmative prescription is just as important as his critical point. He says that we must always ask whether a proposed government power “is necessary to the public good”. Put to one side the question of whether the public good is a useful or meaningful metric. (I have just argued that it is not.) It’s Madison’s necessity standard that I want to emphasize. Necessity is a high bar; it is not enough that a proposed power might be advantageous ― it has to be necessary. This is not obviously a prescription for expansive government, let alone for an expansive administrative state. And then, even with necessary powers, Madison says that we must “guard as effectually as possible against a perversion of the power to the public detriment”. This dovetails nicely with his concern for dispersing and checking powers explored in later (and better-known) papers.

The anti-administrative case isn’t that the administrative state can do no good. Of course it can, sometimes. It is, first, that the administrative state is often actively harmful ― on balance, even accounting for the good it can do ― such that it cannot be regarded as necessary; and, second, that the structure of administrative institutions is such that they fail to provide effectual guardrails against the perversion of their powers. The rest of Law and Leviathan is meant as a response to this last contention and, if its arguent succeeds, it will address part of the anti-administrativists’ concerns. But it will be less important part, as the order of Madison’s requirements makes clear. Devising protections against the abuse of power can only come after we have established that the power is necessary. And anti-administrativists’ concerns on this first front cannot be assuaged by simply pointing to the good that the administrative state might do ― least of all when, as in the example offered by Professors Sunstein and Vermeule ― the good is an illusion that rests on faulty or misleading claims about the nature and effect of coercion in the administrative state and in the market.

Jurisdiction and the Post-Vavilov Supreme Court: Part I

What does “jurisdiction” mean, anyways?

As I wrote in my newsletter last week, the Supreme Court has an awkward relationship with the concept of “jurisdiction.” There is no more tortuous concept in Canadian administrative law. Vavilov, apparently, was the end to the concept of jurisdiction in Canadian administrative law. Vavilov basically said two things about jurisdiction: (1) it is difficult to identify a jurisdictional question, which sheds doubt on the entire enterprise (Vavilov, at para 66); and (2) as a result, “[w]e would cease to recognize jurisdictional questions as a distinct category attracting correctness review” (Vavilov, at para 65). Taken together, it was a fair assumption that jurisdictional questions, if they existed at all, would not be recognized in the law of judicial review.

Easier said than done. The Supreme Court in two recent cases have gone back to the well and drawn from the waters of jurisdiction. In both Ward and Horrocks, the various opinions continue to draw on jurisdiction as a concept without interrogating it. Underneath this technical issue of administrative law is a broader, conceptual difference on the Court that remains post-Vavilov.

In this post I’ll address what I think “jurisdiction” means post-Vavilov. In a future post I’ll address Horrocks and what it might mean for post-Vavilov administrative law splits on the Court.

***

In Ward, under a heading titled, “Jurisdiction Over Defamation and Discrimination,” the majority discusses the “jurisdiction” of the tribunal in that case [28]. In the same paragraph, the Court chastises the Tribunal for indirectly extending its “limited direct jurisdiction.” In Horrocks, on the other hand, the whole dispute concerned the jurisdictional boundary between a labour arbitrator and a human rights tribunal.

The entire setup of these cases is based around the idea of jurisdiction. In Ward, the term was thrown around rather willy-nilly to describe the statutory authority—the grant of power—given to the Tribunal. In Horrocks, the term was used as contemplated by Vavilov, as a category attracting correctness review. But in both cases, jurisdiction looms large.

Before continuing, it’s important to note the various ways that “jurisdiction” has been used in Canadian administrative law. There are at least 3 different uses of the term:

  1. Jurisdiction as a preliminary question: this category concerns “neat and discrete points of law” that arise, for example, in a decision of a human rights commission to refer a case to a human rights tribunal (Halifax, at para 27). In Halifax, the Court overturned previous precedents and held that such questions are reviewable on a reasonableness standard (Halifax, at para 38).
  2. So-called “true questions of jurisdiction”: these questions were said to arise “where the tribunal must explicitly determine whether its statutory grant of power gives it authority to decide a particular matter” (Dunsmuir, at para 59). An example of such a question was provided in Dunsmuir: “whether the City of Calgary was authorized under the relevant municipal acts to enact bylaws limiting the number of taxi plate licences” (Dunsmuir, at para 59). Note, here, that this question trades on the same idea of “jurisdiction” as the preliminary questions doctrine, but there is a difference: ostensibly, this brand of jurisidictional questions concerns an issue that goes to the merits. Vavilov did away with this concept of jurisdictional question, to the extent that such questions attract correctness review.
  3. “Jurisdictional boundaries between two or more tribunals”: this is the category of review at issue in Horrocks. Vavilov retained this category as attracting correctness review.

What is immediately clear is that “jurisdiction” is a morass.

What sense should we make of this? In my view, Vavilov left the door of “jurisdiction” open a crack. The result, as Paul Daly presciently observed the day after Vavilov was rendered, is that jurisdiction is still around—a “stake through the heart” will be the only thing to kill it. In the meantime, we must make sense of what is left of jurisdiction.  As I noted above, one option is to read Vavilov rather broadly: jurisdiction is dead, and we killed it. But this does not explain (in a satisfying way) what the Court is doing in both Ward and Horrocks. Why mention a concept that is dead?

Instead, I think “jurisdiction” (or, as I shall say, hopefully a better label) remains an important concept in Canadian administrative law. This version of jurisdiction—as used in Ward and Horrocks—is not akin to the concept of jurisdiction known to administrative law history (ie) Anisminic. It is not the “preliminary questions” doctrine put to rest in Halifax. This conception of jurisdiction is basically co-extensive with any number of formulations that describes the authority delegated to an administrative decision-maker. The Supreme Court of the United States describes this as “statutory authority,” which is a good a term as any. This is because, fundamentally, any time an administrative decision-maker acts, it is explicitly or implicitly dealing with the boundaries governing it by statute. Whether this is “jurisdiction,” or “statutory authority” does not matter much. It’s all the same thing.

Now, what is true about jurisdiction is that there are different types of legal questions. Some legal questions could be said to be “preliminary.” An example might be a legal condition precedent to the exercise of another legal power under the same statute.  But the difference that Vavilov introduces is simply about the standard of review, not about the existence or not of jurisdictional questions understood in this sense. In other words, to the extent that Halifax and Vavilov dispatched with various types of jurisdictional questions, they only did so to the extent that it matters for the standard of review. Vavilov tells us that questions of jurisdiction, as they were previously known, are hard to identify: and in that sense, they shouldn’t be treated differently than any other legal questions. So whether the question is “preliminary” or on the merits, it’s a legal question that is assimilated to the Vavilov framework.

Why does any of this matter? There is a clarity reason and a substantive reason. For clarity’s sake, the Court should probably not refer to “jurisdiction” anymore. The concept itself, as it is now used, is simply referring to a type of legal question, not a category of review. The Court should adopt some concept of “statutory authority” to describe all the types of legal questions that arise in a typical judicial review proceeding, including anything that might be considered “jurisdictional.” This has nothing to do with the standard of review: all of the questions will be presumed to be reviewed on reasonableness review. On the substantive side, and as we shall see from Horrocks, there are good reasons to take statutes—and the boundaries they set up—seriously. As Vavilov says, the discarding of jurisdiction as a category of review should not lead to  the arrogation of administrative power.

It Ends Well

Thoughts on the Supreme Court’s narrow but seemingly decisive rejection of a right not to be offended

Last week, the Supreme Court delivered its judgment in Ward v Quebec (Commission des droits de la personne et des droits de la jeunesse), 2021 SCC 43. By a 5-4 majority, it quashed an award of damages a human rights tribunal had granted to Jérémy Gabriel, a child celebrity, whom a well-known comedian, Mike Ward, had cruelly mocked. As Jen Gerson and Matt Gurney put it in The Line’s editorial (possibly paywalled, but you should subscribe!)

Ward … decided to become That Asshole, the edgelord comedian who pointed out that the kid wasn’t very good. In a few stand-up bits, Ward called the child ugly, and noted that the performances were tolerable only because he thought the singer’s condition was terminal. Nice guy. (Paragraph break removed)

The tribunal, and the Québec Court of Appeal found that this amounted to discrimination in the exercise of Mr. Gabriel’s right to “the safeguard of his dignity” under section 4 of Québec’s Charter of Human Rights and Freedoms, a.k.a. the Québec Charter. The majority of the Supreme Court resoundingly holds otherwise.

Instead of my usual blow-by-blow summary and comment, I will offer some more condensed thoughts on a few striking aspects of this case. While the most important thing about Ward is what, if anything, it means for the freedom of expression, there are a few other things to mention before I get to that. In this post, I mostly focus on the majority opinion. I will shortly post separately about the dissent.

The Human Face

Because I will argue that the majority decision is correct, and indeed that it was very important that Mr. Gabriel not win this case, I want to start by acknowledging that he has had it very hard. Mr. Ward’s jokes at his expense were cruel. Mr. Gabriel did suffer, greatly ― we are told that he even tried to kill himself at one point. I think we can wonder whether the connection between these things is all that strong. I’m not persuaded by the dissent’s imputation to Mr. Ward of the full responsibility for Mr. Gabriel’s bullying by his classmates. We can also argue that anti-discrimination law ― perhaps any law ― isn’t the solution. But we have to recognize that a person has been in a lot of undeserved pain, and a person who, even before this case, had not had it easy in life.

The Court

As already noted, the Court is narrowly divided. The Chief Justice and Justice Côté write for the majority, with Justices Moldaver, Brown, and Rowe concurring. Justices Abella and Kasirer write for the dissent, joined by Justices Karakatsanis and Martin. For those keeping score at home, this is the exact same alignment as in the recent decision in Toronto (City) v Ontario (Attorney General), 2021 SCC 34. Indeed, even the authorship of the opinions overlaps: in City of Toronto, the Chief Justice wrote with Justice Brown, while Justice Abella wrote for the dissenters.

I’m old enough to remember, as they say, how smugly self-satisfied Canadian commentators were, just a few years ago, at the consensus reigning at our Supreme Court, in contrast to the US one always splitting 5-4. To be sure, two cases do not make a trend, but I think it’s pretty clear that on the Supreme Court as it has recently been constituted there is ― though there are always exceptions ― a somewhat cohesive group consisting of Justices Côté, Brown, and Rowe, and perhaps an even more cohesive group led by Justice Abella, with Justices Karakatsanis, Martin, and Kasirer. The Chief Justice and Justice Moldaver are the swing votes. It remains to be seen how, if at all, Justice Abella’s retirement is changing this, but in the meantime, our Supreme Court has been fractured along lines that can be predicted. This is not necessarily bad. But let’s not be smug.

One odd thing to add is that, whereas in City of Toronto majority and dissent were ― by the standards of the Supreme Court of Canada ― at each other’s throats, here they studiously ignore one another. I’m not sure which is better, but the contrast between cases argued and decided just a month apart, by identical alignments, and with overlapping opinion authorships, is striking.

The Case

One uncomfortable question I have is: should the Supreme Court have taken this case at all? Let me take you straight away almost to the end of the majority judgment, where we learn, for the first time, the following

[I]n light of the Tribunal’s finding that Mr. Ward [translation] “did not choose Jérémy because of his handicap” but rather “because he was a public personality” (Tribunal reasons, at para. 86), it must be concluded that the distinction was not based on a prohibited ground. This conclusion on its own is sufficient to dispose of the appeal. [91]

Everything else that the Court has said and that I’m about to discuss ― that’s just obiter dicta. The tribunal made a basic logical mistake, which, as the majority explains, the Court of Appeal then glossed over. That was, of course, unfortunate. But it’s not the Supreme Court’s role to correct basic logical mistakes by tribunals or even courts of appeal. They’re there to develop the law. And develop the law they do ― in a way that, if the majority is right (and I think it is), was pressing and necessary. But also in a way that, by the majority’s own admission, is beside the point in this case.

I think this raises the issue of the Supreme Court’s role in our constitutional system. Where is the line between developing the law in deciding cases, as we expect them to, and developing the law by making big pronouncements that are unnecessary to decide cases? Should a court refrain from doing the latter, or may it properly seize on the opportunities that present itself to it to provide important guidance to lower courts? I have no firm views on any of this, but I think the questions are worth thinking about. (For some related musings, see here.)

Jurisdiction

Back to the very beginning of the majority’s reasons:

This appeal … invites us … to clarify the scope of the jurisdiction of the Commission des droits de la personne et des droits de la jeunesse … and the Human Rights Tribunal … with respect to discrimination claims based on the … Quebec Charter. [1]

Clarify the… what? Yes. That word. The majority uses it several times in the course of its reasons. In particular, it speaks of “the distinction that must be drawn with respect to jurisdiction over, on the one hand, an action in defamation and, on the other, a discrimination claim in the context of the Quebec Charter“. [22]

This is odd. A mere two years ago, in Canada (Minister of Citizenship and Immigration) v. Vavilov, 2019 SCC 65, all of the Ward majority judges signed an opinion that not only eliminated jurisdictional questions as a distinct category of correctness review, but seemed to endorse scepticism at the very “concept of ‘jurisdiction’ in the administrative law context”. [66] Vavilov said that what might previously have been thought of as jurisdictional questions are legal questions like all others, subject to reasonableness review, except when the respective jurisdictions of two administrative bodies must be demarcated.

One recent example of this reasoning is the decision of the Ontario’s Superior Court of Justice in Morningstar v WSIAT, 2021 ONSC 5576, about which I have written here. The Court roundly rejected the argument that, as I summarized it

the jurisdictional boundary between a tribunal and the ordinary courts should be policed in much the same way as, Vavilov said, “the jurisdictional boundaries between two or more administrative bodies”, [63] ― that is, by hav[ing] the court ensure the boundary is drawn correctly.

I thought ― and still think ― that that was a correct application of Vavilov. Ward, though, says that there is indeed a jurisdictional boundary between administrative tribunals and courts. I don’t think this is consistent with Vavilov. Nothing turns on this here because the case gets to the courts by way of statutory appeal rather than judicial review, and ― under Vavilov ― the correctness standard applies to all legal questions in such circumstances. But the tensions inherent in Vavilov, including in its attempt to rid Canadian administrative law of the fundamental concept of the law of judicial review are becoming apparent. (Co-blogger Mark Mancini has made a similar observation in the latest issue of his newsletter.)

Interpretation

One of the things the majority is right about is that Ward is, among other things, a case about interpretation. It requires the courts to make sense of a somewhat peculiar statutory scheme, which protects, among other things, rights to the freedom of expression and to the “safeguard of [one’s] dignity”, says that “the scope of the freedoms and rights, and limits to their exercise, may be fixed by law”, and protects equality in “the exercise and recognition” of these rights, rather than as a general self-standing right. This is not an easy exercise and I won’t go into all the details, but I will make a few comments.

The majority deserves credit for trying to work out an independent meaning for the right to the safeguard of one’s dignity. As it notes, dignity is a very tricky concept ― and the Supreme Court itself has tried to avoid putting too much weight on it in other contexts. But here it is, in the text of the Québec Charter, a statute that binds the courts. It will not to do to simply find violations of dignity when other rights are violated in particularly egregious ways, as Québec courts had done. The Québec Charter makes it a distinct right, and the courts must treat it as such. At the same time, they have to give it defined contours. The majority seeks to do so by stressing the importance of the safeguard of dignity, to which the right is directed:

Unlike, for example, s. 5 [of the Québec Charter], which confers a right to respect for one’s private life, s. 4 does not permit a person to claim respect for their dignity, but only the safeguarding of their dignity, that is, protection from the denial of their worth as a human being. Where a person is stripped of their humanity by being subjected to treatment that debases, subjugates, objectifies, humiliates or degrades them, there is no question that their dignity is violated. In this sense, the right to the safeguard of dignity is a shield against this type of interference that does no less than outrage the conscience of society. [58]

What the majority does is a careful and, I think, pretty convincing reading of the statutory text. Good.

Some things the majority says are not so good. For instance: “the interpretation of this provision must be refocused on its purpose by considering its wording and context”. [55] No, no, no. Interpretation should be focused on text understood in context. Purpose can sometimes help a court understand the words and enrich its understanding of the context, but it should not be the focus of interpretation. And then, there is this:

This Court’s jurisprudence also establishes “that mere differences in terminology do not support a conclusion that there are fundamental differences in the objectives of human rights statutes” … It follows that, as long as this is not contrary to the usual rules of interpretation, symmetry in the interpretation of the various instruments that protect human rights and freedoms is desirable. [68; quoting Quebec (Commission des droits de la personne et des droits de la jeunesse) v Montréal (City), 2000 SCC 27, [2000] 1 SCR 665, [47]]

What are we to make of this? If usual rules interpretation are to prevail, differences in terminology must make a difference, if not to the objectives then to the effects of human rights as of any other statutes. And the idea that differences in wording don’t matter because objectives are key to interpretation is specifically rejected in the majority opinion in Quebec (Attorney General) v 9147-0732 Québec inc, 2020 SCC 32 ― signed onto by the same five judges who are in the majority in Ward (even as it is endorsed by the concurrence).

Between the jurisdiction issue and this, I cannot help but wonder whether their Lordships remember what they said last year. Or are they trying to say that we are supposed not to? This stream of inconsistent pronouncements ― by the same people! ― reveals, at best, a lack of attention to legal doctrine and craft. It is very disappointing.

Freedom of Expression

I finally come to the meat of the case. Here too, I want to praise the majority for getting things fundamentally right, but also to criticize them for saying things along the way that are doubtful or even wrong in themselves, or inconsistent ― without explanation ― with important precedent.

Let me start with a quick note from the “judges are not philosophers” file. The majority’s discussion of the freedom of expression begins with the assertion that it, “[l]ike the right to the safeguard of dignity … flows from the concept of human dignity”. [59] Perhaps. But in the next paragraph the majority quotes Joseph Raz’s claim that “a person’s right to free expression is protected not in order to protect him, but in order to protect a public good, a benefit which respect for the right of free expression brings to all those who live in the society in which it is respected”. [60] These are two quite different views of the foundations and purposes of the freedom of expression ― one deontological, the other utilitarian. Perhaps nothing turns on which of these is correct in this case, but if so, the majority shouldn’t be making these philosophical declarations at all. And I suspect that in some cases the choice might actually make a difference. The majority’s approach is muddled and unhelpful.

Now for some good things. This, especially: “freedom of expression does not truly begin until it gives rise to a duty to tolerate what other people say”. [60] This is the key to so many disputes about freedom of expression. Speech is not harmless. It can hurt. It can propagate falsehoods. It can inflame base passions. But freedom of expression means sometimes having to tolerate such things ― just like freedom of assembly means having to tolerate noisy protests, and freedom of religion means having to tolerate heresy and blasphemy ― even when their cost falls on particular groups or even individuals.

The majority adds that “[l]imits on freedom of expression are justified where, in a given context, there are serious reasons to fear harm that is sufficiently specific and cannot be prevented by the discernment and critical judgment of the audience”. [61] This sets a fairly high bar to limits that will be considered justified. It also acknowledges that the audience has its share of responsibility in appreciating troublesome words. Courts assessing a limit on the freedom of expression should not assume that citizens are, by default, unthinking and gullible playthings for the tellers of tall tales. This is also good and important. Assuming away all critical sense among the citizens would help justify all kinds of restrictions on speech, including, and perhaps especially, in the political arena. It is fundamentally incompatible with the notion of a self-governing, responsible citizenry.

But this insistence sits uneasily, to say the least, with the Court’s position in Harper v Canada (Attorney General), 2004 SCC 33, [2004] 1 SCR 827. There, the majority said that

The legislature is not required to provide scientific proof based on concrete evidence of the problem it seeks to address in every case. Where the court is faced with inconclusive or competing social science evidence relating the harm to the legislature’s measures, the court may rely on a reasoned apprehension of that harm. [77]

This is contrast to the Harper dissent’s concern that “[t]here [was] no demonstration that” the limits on “third party” spending at issue were “required to meet the perceived dangers of inequality, an uninformed electorate and the public perception that the system is unfair”. [38] By my lights, Ward‘s insistence on serious reasons to fear specific harm, as well as on audience discernment is much more in tune with the Harper dissent. Because I regard Harper as an abominable decision, I am happy to see Ward go in a different direction. But there is no comment in Ward on how these cases interact. Again, it’s as if the judges don’t remember what the law says, though at least Harper is a much older case that Vavilov and Québec Inc.  

All that said, the substance of the majority’s decision is right and reassuring (or it would be reassuring if more than five judges had signed on). The majority insists that the right to the safeguard of one’s dignity most not be “vague” or given “a scope so broad that it would neutralize freedom of expression”. [80] It stresses the objective nature of the test for whether this right is breached and rejects the modified objective standard of “a reasonable person targeted by the same words”, because “[t]hat approach results in a shift toward protecting a right not to be offended, which has no place in a democratic society”. [82] What matters is neither “the repugnant or offensive nature of the expression [nor] the emotional harm caused”, [82] but the effect of the words on listeners: would “a reasonable person, aware of the relevant context and circumstances, … view the expression … as inciting others to vilify [its targets] or to detest their humanity on the basis of a prohibited ground of discrimination” [83] and would “a reasonable person would view the expression, considered in its context, as likely to lead to discriminatory treatment of the person targeted”? [84]

All this is the more important since the list of prohibited grounds of discrimination under the Québec Charter is very broad and includes “political convictions”. As I have written here, “even if we accept the need to protect vulnerable minorities from hate speech targeting them, I struggle to see what makes it necessary to extend this protection to members of political parties or movements”. Protecting people from mockery, let alone hurt feelings, based on their political views is incompatible with lively democratic debate. However much we can wish for such debate to usually be civil, I think it’s a mistake to insist that it always must be, and certainly a grave mistake to put government officials in charge of deciding whether it is sufficiently civil on any give occasion.


The insistence on the need for objective assessment and the clear rejection of a right not to be offended will, I hope, be the key takeaway from Ward. For them, we can forgive the majority opinion its many flaws. That there can be no right not to be offended in a society that proclaims its commitment to the freedom of expression and to democracy might have been self-evident ten years ago, but it evidently isn’t anymore. The dissent offers us a glimpse of what a world in which this truth isn’t recognized looks like. I will focus on it in a forthcoming post.

What Does City of Toronto Mean For Administrative Law?

The Supreme Court released its much-anticipated decision today in Toronto (City) v Ontario (Attorney General), 2021 SCC 34. While others will address the nuances of the case, the majority generally puts unwritten constitutional principles into a tiny, little box. It says that because “[u]nwritten principles are…part of the law of our Constitution…” [50], unwritten principles only have two practical functions: (1) they can be used in the interpretation of constitutional provisions [55]; (2) they can be used to “develop structural doctrines unstated in the written Constitution per se, but necessary to the coherence of, and flowing by implication from, its architecure” [56]. In this category, the Court uses the example of the doctrine of paramountcy, the doctrine of full faith and credit, and the remedy of suspended declarations of invalidity.

I applaud the majority opinion for clarifying the role of unwritten constitutional principles. For my part, I think the functions they have outlined for unwritten principles give those principles a meaningful role in the constitutional structure while giving priority to the text. The majority aptly underscores the worry with unwritten principles–they are so abstract and potentially endless–and negates that worry by ensuring the text as a control on the use of these principles. Even better, the majority closes the door on the rather pernicious attempt to read municipalities into s.3 of the Charter [5].

But that is not my concern for today. What does any of this have to do with administrative law?

Post-Vavilov, there was a good argument that unwritten principles–the Rule of Law specifically–could have independent force in limiting state action in some way on the standard of review–put more bluntly, that the Rule of Law could invalidate certain legislative rules governing standard of review. The Court says, for example, that “where the legislature has indicated the applicable standard of review, courts are bound to respect that designation, within the limits imposed by the rule of law” (Vavilov, at para 35). It goes on to outline categories of questions–like constitutional questions–that demand a correctness standard because of “respect for the rule of law” (Vavilov, at para 53). This raised the argument that if a legislature were to prescribe a standard of review of reasonableness on a constitutional question, such a standard would not be given effect to by a court because it transgresseses the “limits imposed by the rule of law.”

On first blush, City of Toronto tends to throw cold water on the argument. Its insistence that unwritten principles cannot invalidate legislation could mean that a court should give effect to a legislated standard of review on constitutional questions. And because there is no express constitutional provision insisting on a correctness standard on certain questions, on a strict reading of the City of Toronto majority opinion, there would be no power to invalidate that law.

This very well may be true, and yet I think there are a few ways to reconcile City of Toronto with Vavilov that leads to the same result that Vavilov seems to suggest–a court not applying (which is strictly, though perhaps not functionally, different from invalidation) a legislated standard of review of reasonableness on constitutional questions. Much of this argument hinges on s.96 of the Constitution Act, 1867.

First, it might be said that the Rule of Law as outlined in Vavilov is a necessary interpretive principle that should be used to understand s.96. That is, we cannot understand s.96–which contemplates federally-appointed superior courts–without understanding the traditional role of these courts to conduct judicial review of administrative action on a certain stringency on certain questions. In City of Toronto, the Court cites s.96-100 as an example of unwritten principles bolstering a constitutional principle, suggesting that “unwritten constitutional principles of judicial independence and the rule have law have aided in the interpretation of [ss.96-100], which have come to safeguard the core jurisdiction of the courts that fall within the scope of those provisions” [55].

I think to call any of the doctrinal innovations that have come to s.96 a result of “interpretation” stretches the term a bit far. On its face, s.96 is just an appointing provision. It may be one thing to interpret what the terms of that appointing provision are, but to construct doctrine on top of the provision–or to make it work in a constitutional structure–seems to be a different judicial function.

Secondly, and I think more persuasively, the Court notes that unwritten principles can develop structural doctrines that flow from constitutional architecture [56]. Again, the Court notes examples of this sort of doctrinal construction: full faith and credit, paramountcy, and even the legal result in the Quebec Secession Reference. As we see, some of these doctrines are quite particular to specific contexts–the Quebec Seccession Reference, for example. Others are more general. The doctrine of full faith and credit in the context of conflict of laws is a major doctrinal innovation that is not found anywhere in a specific constitutional provision. These doctrinal innovations can, in effect, change or invalidate legislation that conflict with them, though they are rooted in the text itself.

Vavilov‘s comments on standard of review best fall into this category. The standard of review framework flows from two unwritten principles themselves: legislative intent (perhaps partially reflected in the principle of “democracy”) and the Rule of Law. The Court conceives of the Rule of Law as generally the rule of courts, in that courts must retain a strong supervisory role over certain questions. It would upset the supervisory role of these courts to outlaw their ability to hold state actors to the strictest constitutional standard. This is but a logical extension of Crevier, which set the stage for an argument about the constitutionally-protected role of the superior courts.

An example and a caveat. First, the majority and dissent clash over MacMillan Bloedel. In that case, the Court arguably invalidated a legislative scheme that granted exclusion jurisdiction to a youth court. The City of Toronto majority says the holding in that case was based on the text of ss.96-101 and 129 of the Constitution Act, 1867 [50]. The dissent, on the other hand, cites para 41 of MacMillan Bloedel to suggest that the basis of the holding was the Rule of Law itself [176]. In my view, MacMillan Bloedel is a bit of both. The Court clearly bases its decision in s.96 (MacMillan Bloedel, at para 47). But it also says that the case is best understood “in a broader constitutional context, considering this jurisprudence along with the preamble to the Constitution Act, 1867, the principle of the rule of law, and the central place of superior courts in our system of governance” (MacMillan Bloedel, at para 2). To the extent these principles and s.96 were abridged, the impugned legislative provision was “read down” as “inoperative to deprive the superior court of its jurisdiction to convict the appellant of contempt in this case” (MacMillan Bloedel, at para 43). In MacMillan Bloedel, we have a constitutional text (s.96)–>supported by the Rule of Law (unwritten principle)–>a result that the core of superior court powers were protected in this case. Vavilov falls into this same category. We can see, then, that in some cases a legislative standard of review may be “read down” as a result of the standard of review doctrine spun out from the unwritten principles of legislative intent and the Rule of Law.

The caveat I wish to raise has to do with the Federal Courts. Section 96 does not speak to statutory courts, and in theory, the Federal Courts’ judicial review jurisdiction could be abolished tomorrow unlike the superior courts. All of this, then, would stop at the Federal Courts. But I do not think this is inevitable. Once a statutory court has been made under s.101 of the Constitution Act, 1867, one might make the argument that so long as such a court exists, its powers should be construed as broadly as the powers of a superior court under s.96. But I do not commit to this argument in full, except to say that it makes practical sense to me and would uphold a consistent judicial standard for administrative action across jurisdictions.

At any rate, I think City of Toronto–despite its strong language on unwritten principles–can be reconciled with Vavilov. And at the end of the day, the result may be the same: legislation that undermines an unwritten principle may not be “given effect” according to a doctrinal innovation, even if the legislation is not “invalidated” in a strict sense. This is the best way to undertstand Vavilov‘s standard of review framework.