Vavilov On the Road
In the first year after Vavilov was released, I read pretty well every decision in which the Supreme Court of Canada’s reformulation of administrative law was cited. The result was “One Year of Vavilov” and a soon-to-appear chapter in Colleen Flood and Paul Daly eds., Administrative Law in Context, 4th ed.
Since then, I have been less assiduous in reviewing the post-Vavilov case law — though I still skim the vast majority of new decisions — but that is because there is less to take specific note of. Interesting cases and novel issues crop up now and then, but broadly speaking the jurisprudence is stable, with lower courts faithfully applying the Vavilov framework, as indeed they have been since December 2019. The direction of travel is settled.
In this post, I want to draw attention to cases and issues which have caught my attention so far this year.
Selecting the Standard of Review
The Vavilovian starting point is, of course, a presumption of reasonableness review, subject to exceptions based on institutional design and the rule of law. The institutional design exceptions include, notably, the application of appellate standards of review where a statutory right of appeal has been created (see Silver and Federer, (2021) 34 CJALP 179). Correctness review applies only to extricable questions of law and, so far, courts have been relatively parsimonious in identifying such questions, especially in professional discipline cases (see e.g. Byun v. Alberta Dental Association and College, 2021 ABCA 272). Where concepts familiar to common-law courts are in play, however, appellate oversight can be expected to be rigorous (see e.g. Partridge v. Nova Scotia (Attorney General), 2021 NSCA 60, at para. 69 (misapplication of the test for causation by the Utility and Review Board was an extricable error); Quadra Properties Ltd. v Gamble, 2021 SKQB 16 (tenancy agreement)).
The most significant post-Vavilov decision on the rule of law exception to the presumption of reasonableness review is undoubtedly United Nurses of Alberta v. Alberta Health Services, 2021 ABCA 194. (Disclosure: I am counsel for Alberta Health Services in its application for leave to appeal to the Supreme Court of Canada.) The underlying issue in this labour relations matter was the test for making out a prima facie case for discrimination on the basis of family status (on which see this comprehensive treatment by Vipond and Oliphant). A majority of the arbitration board had applied the Federal Court of Appeal’s Johnstone test and rejected the grievor’s discrimination claim. But the first-instance judge and the Court of Appeal rejected the Johnstone test. For the Court of Appeal, the quasi-constitutional status of human rights legislation – the backdrop against which the collective agreement here fell to be interpreted – meant that correctness simply had to be the applicable standard of review as a ‘general question of law of central importance to the legal system’ arose:
The interpretation of the same human rights protections in collective agreements and in human rights legislation must be consistent, as these provisions provide some of the most important protections in our society. Labour arbitration boards, human rights tribunals, and superior courts on review are regularly called upon to consider discrimination cases and to interpret similar or identical human rights protections in keeping with human rights legislation, which has quasi-constitutional status in Canada (at para. 55).
Even allowing for my professional interest in this matter, I think it is fair to say that this is as broad a view of the scope of this correctness category as has been taken post-Vavilov: contrast, e.g. Gaudreau c. Régie des marchés agricoles et alimentaires du Québec, 2021 QCCA 330, at paras. 31-32.
Most of the time, however, reasonableness is the appropriate standard of review. And in its “essential nature”, it remains a deferential standard (Ontario Nurses’ Association v. Participating Nursing Homes, 2021 ONCA 148, at para. 104, per Huscroft JA, dissenting). Ruel JA provided a sharp reminder of this in M.O. c. Société de l’assurance automobile du Québec, 2021 QCCA 177, criticizing counsel for essentially repeating arguments made not just before the reviewing court but also before multiple administrative decision-makers:
La multiplication des recours en proposant, palier après palier, année après année, les mêmes arguments que ceux initialement plaidés et évalués par des instances administratives spécialisées, sans tenir compte des préceptes fondamentaux du droit administratif, est de nature à frustrer les justiciables, qui ne se voient pas offrir la justice administrative souple et efficace à laquelle ils sont en droit de s’attendre (at para. 66).
Administrative decisions are to be read fairly; reviewing courts are not to prowl the reasons and record looking for any small error on which they can pounce (Alexion Pharmaceuticals Inc. v. Canada (Attorney General), 2021 FCA 157, at paras. 12-18):
In Vavilov, the Supreme Court tells us that we should not be too hasty to find [material] flaws. Vavilov’s requirement of a reasoned explanation cannot be applied in a way that transforms reasonableness review into correctness review. If reviewing courts are too fussy and adopt the attitude of a literary critic all too willing to find shortcomings, they will be conducting correctness review, not reasonableness review. That would return us to the bad old days in the 1960’s and 1970’s when reviewing courts would come up with any old excuse to strike down decisions they disliked—and often did (Canada (Citizenship and Immigration) v. Mason, 2021 FCA 156, at para. 37, per Stratas JA. See also Canada (Attorney General) v. Association of Justice Counsel, 2021 FCA 37, at para. 10; Manitoba Government and General Employees’ Union v. The Minister of Finance for the Government, 2021 MBCA 36, at para. 59; Trentway-Wagar inc. c. Cormier, 2021 QCCA 983, at para. 35).
That said, Vavilovian reasonableness review contains requirements of justification, responsiveness and contemporaneity that demand more than was previously demanded of decision-makers. Decision-makers who were already implementing best adjudicative practices prior to Vavilov should continue to see their decisions upheld: see e.g. Amalgamated Transit Union, Local 615 v. Saskatoon (City), 2021 SKCA 93, at paras. 98-120; Gezehegn v. Alberta (Appeals Commission of the Workers’ Compensation Board), 2021 ABCA 93, at paras. 12-23; Xiao v. Canada (Citizenship and Immigration), 2021 FC 386, at para. 30 (unless, of course, only one outcome is possible because of the constraints of statutory language, as (the majority found) in Ontario Nurses’ Association v. Participating Nursing Homes, 2021 ONCA 148 or the evidence in the record, as in A c. Procureur général du Québec, 2021 QCCA 599, at paras. 26-32). But decision-makers who benefited in an earlier era from reflexive deference on the part of courts are having to up their game (Syndicat national des convoyeur(e)s de fonds (SNCF – SCFP), section locale 3812 c. Hamelin, 2021 QCCS 932, at para. 34; Glencore Canada Corporation c. Syndicat des métallos, section locale 9449, 2021 QCCS 3357, at paras. 36-37).
An important ministerial decision failed to pass muster in Manitoba Government and General Employees’ Union v The Minister of Finance for the Government, 2021 MBCA 36 (see also Northern Harvest Smolt Ltd. v Salmonid Association of Eastern Newfoundland, 2021 NLCA 26. Cf Canada (Health) v. Glaxosmithkline Biologicals S.A., 2021 FCA 71. The issue here was the Minister’s refusal, in the context of a dispute about the negotiation of a collective agreement, to appoint an arbitration board at the request of the Union, made under s. 48 of the Civil Service Act, CCSM c C110. The Minister’s view was that he retained “discretion to decide whether or not there were ‘matters on which agreement [could not] be reached’” within the meaning of s. 48 and “he interpreted this phrase as requiring the presence of an objective bargaining impasse” (at para. 67). The reasons provided by the Minister — in letters to the Union — were insufficient:
However, there is nothing in the Minister’s reasons or the history and context of the matter before him that demonstrates why there was such cursory treatment of this key issue. For example, there is no reference to precedents supporting a Minister’s discretion in this particular context, or to an applicable public interpretive policy (see para 94). Furthermore, regarding the history and context of the matter, as explained above, the MGEU specifically contested the existence of the Minister’s purported exercise of discretion. There was nothing in the Minister’s reasons, nor was there any other contextual circumstances, which suggested that his cursory, conclusory response was reasonable (at para. 75).
Furthermore, Steel JA explained, the legislation (when considered in French and English) was “clear” that it is the requesting party which identifies the matters upon which agreement cannot be reached, not the Minister (at para. 82); there is “no general discretion to refuse to appoint an arbitration board” once the statutory conditions are satisfied (at para. 91). (One might say that this was a “pouvoir lié”, albeit that the question is always whether the Minister’s interpretation is reasonable: for discussion, see 11316753 Canada Association c. Canada (Transports), 2021 FC 819, at para. 30.)
Another example is drawn from the labour relations context: Bragg Communications Inc. v. Unifor, 2021 FCA 59. In keeping with its practice of not writing detailed reasons to support certain orders, the Canada Industrial Relations Board granted an application for the expansion of a bargaining unit “in a concise order without accompanying reasons” (at para. 4). This was insufficient, as Near JA observed:
Without the benefit of any of its reasoning on the issue of appropriateness of the bargaining unit, it is impossible to assess whether its conclusion is reasonable. This does not mean that separate, detailed formal reasons were required. It simply means that the Board was required to set out its reasoning on the issue of appropriateness, just as it did regarding the question of the union’s representative character. The respondent union in its submissions sought to fill this void but it is the responsibility of the Board to set out the rationale for its decision and failure to do so in this case makes the decision unreasonable (at para. 10).
Lastly, front-line decision-makers (especially those with high volumes of matters to address) continue, at least on occasion, to find the hurdles of justification, responsiveness and contemporaneity too high to scale. For example, in Zhang v. Canada (Public Safety and Emergency Preparedness), 2021 FC 746, at para. 31, a decision was based on a report which in turn was based on post-dated findings and, thus, unreasonable. (See alsoParty A v. The Law Society of British Columbia, 2021 BCCA 130, at para. 30 (professional regulation); Al-Naami v. College of Physicians and Surgeons of Alberta, 2021 ABQB 549, at para. 205 (professional regulation); Lesko v. Attorney General of Canada, 2021 ONSC 2883, at paras. 53, 59 (prison); Green v. Canada (Attorney General), 2021 FC 178, at paras. 41, 46 (Treasury Board official); Centre intégré de santé et de services sociaux de Laval c. Brassard, 2021 QCCS 406 (labour arbitrator); Anku v. Canada (Citizenship and Immigration), 2021 FC 125, at para. 31(senior immigration officer)).
It is unlikely, however, that those who have failed to scale Vavilov’s slightly higher bar will not try harder in the future. All things being equal, over time, decision-makers will come to heed the Supreme Court’s call for justification, responsiveness and contemporaneity. Notably, in Beaudoin v. British Columbia, 2021 BCSC 512, at paras. 232-246, a series of public health measures taken to combat the pandemic were upheld because the Provincial Health Officer made a reasoned decision based on evidence which was, moreover, sufficiently alert, alive and sensitive to the impact of the measures on fundamental rights and freedoms. Would her reasons have been so extensive before Vavilov? Perhaps not: if so, this is an example of the reflexive effect of judicial review doctrine on decision-making on the front lines of public administration.
Reasonableness Review, Procedural Fairness and Appellate Tribunals
One narrow issue which has cropped up on occasion involves appellate administrative tribunals and procedural fairness. Where a first-instance decision-maker (like a faculty committee at a university, or the Refugee Protection Division) allegedly breaches the duty of fairness, and the appellate decision-maker (like a university committee, or the Refugee Appeal Division) considers that the duty was not breached, what is the standard of review? Should the reviewing court simply apply the Baker factors or should it apply the reasonableness standard to the reasons the appellate decision-maker provided? As McHaffie J observed in Abiodun v. Canada (Citizenship and Immigration), 2021 FC 642, the fact that reasons are provided does not transform an issue of fairness into an issue of substance: reasons can be a useful proxy for substance in difficult cases, but a procedural fairness issue at first instance remains a procedural fairness on appeal. Nonetheless, at the very least, where a procedural fairness determination “is in turn based on factual findings or an assessment of the evidence, those findings themselves continue to warrant deference” (at para. 10). Indeed, in other cases judges have applied the reasonableness standard in reviewing an appellate decision-maker’s express consideration of a procedural fairness issue: see e.g. Ibrahim v. Canada (Citizenship and Immigration), 2020 FC 1148, at paras. 11-18; Andres v. University of Lethbridge, 2021 ABQB 551, at para. 29. Perhaps the best answer, given the injunction in Vavilov to keep procedure and substance distinct, is simply to give the decision-maker who has considered a procedural fairness issue a “margin of deference” (Mission Institution v. Khela, 2014 SCC 24,  1 SCR 502, at para. 89).
Limited Rights of Appeal
One obvious post-Vavilov flashpoint relates to limited rights of appeal. The long-standing practice, certainly at the Federal Court of Appeal and Ontario’s Divisional Court, has been to take limited rights of appeal seriously: if legislation says matters may be appealed only on questions of law or jurisdiction, then an appellant cannot raise issues of fact, policy or discretion on appeal. But paragraph 52 of Vavilov clearly states that matters falling outside a limited appeal clause can nonetheless be judicially reviewed, presumably (though this was left unclear) by way of an application for judicial review and an appeal being initiated concurrently.
There are now three detailed analyses of the implications of paragraph 52, one from Ontario and two from the Federal Court of Appeal, following different analytical frameworks and ending up in different places.
In Yatar v. TD Insurance Meloche Monnex, 2021 ONSC 2507, the Divisional Court refused to entertain a concurrent appeal and application for judicial review of a decision of the Licence Appeal Tribunal refusing an application for statutory accident benefits. For Kristjanson J, the appeal could not be entertained, as it raised questions of mixed fact and law falling outside the scope of the appeal on questions of law only. And the judicial review application should not be entertained: judicial review is discretionary and, where there is a right of appeal, should be entertained only in exceptional circumstances. She gave four reasons justifying the refusal to entertain the judicial review application.
First, the legislature had plainly intended to limit oversight of factual matters in the statutory accident benefits field, implementing a suite of reforms “designed to provide a streamlined response, prioritizing access to justice in a quicker and more efficient manner” (at para. 41). (It is notable, in this regard, that the applicant’s claim was based on an accident which occurred in 2010 and that her engagement with the statutory accident benefits administrative machinery dated to 2012.) Second, there is an internal reconsideration power, exercisable on a basis “akin” to the correctness standard (at para. 43). Third, the nature of the alleged errors — on questions of fact or mixed questions of law and fact involving the assessment of evidence — was such that any judicial review would be conducted on a “high standard of deference” (at para. 44). Fourth, concurrent appeals and judicial reviews create “systemic difficulties”:
The concurrent pursuit of two remedies has triggered two sets of procedures and the filing of voluminous materials. Ms. Yatar filed both a judicial review record and a separate appeal record, due to different requirements under the Rules of Civil Procedure R.R.O. 1990, Reg. 194. The parties each filed two factums, addressing different issues and different standards of review. The LAT has a right to participate on the judicial review, but not on the appeal unless leave is granted. The time periods set out in the Rules of Civil Procedure for filing materials vary between appeals and applications for judicial review. The duplication of materials is a heavy burden on the parties and the court in terms of time, cost and efficiency. These concerns only increase when an application for judicial review is scheduled after an unsuccessful statutory appeal in order to avoid allegations of prematurity (at para. 45).
Accordingly, judicial review would only be available in “exceptional circumstances” which were not present here (at para. 46).
The majority of the Federal Court of Appeal approached the matter quite differently in Canada (Attorney General) v. Best Buy Canada Ltd., 2021 FCA 161. At issue here was a decision of the Canadian International Trade Tribunal on the tariff classification of items Best Buy wishes to import: are television stands “furniture” or are they “parts” of televisions?
Under the Customs Act, RSC 1985, c 1 (2nd Supp), there is a right of appeal to the Tribunal on questions of law: s. 68(1). A privative clause functions as a deadbolt blocking any recourse to judicial review: s. 67(3). Unlike in some other federal regulatory regimes, such as telecommunications or railways, there is no right of appeal to the federal cabinet. Here, there was no extricable question of law: the Tribunal’s refusal to follow the World Customs Organization’s classification of the stands at issue involved a fact-sensitive assessment of a variety of matters, including the WCO’s classification (at para. 24). Could Canada – dissatisfied with the Tribunal’s decision — nonetheless attack factual determinations or determinations of mixed fact and law before the Federal Court of Appeal?
Near JA thought not. He leaned heavily on the Supreme Court’s reliance on “institutional design” in Vavilov. If legislative intent is to be taken seriously, he reasoned, Parliament’s considered choice to restrict appellate oversight should be respected: “If Parliament’s institutional design choices are to be respected, factual issues and issues of mixed fact and law for which no legal question can be extracted must not be subject to review by this Court” (at para. 46). Near JA did not see any constitutional problems arising as a result of limiting appeals from the Tribunal to extricable questions of law, as the letter and spirit of the Supreme Court’s seminal decision in Crevier v. A.G. (Québec) et al.,  2 SCR 220 (to the effect that there is a core constitutional minimum of judicial review) were respected by the scheme of the Customs Act:
The Supreme Court in Crevier was pre-occupied with the lack of any appeal from a decision of the Professions Tribunal to a Superior Court. Comparing to the legislative scheme at issue in this case, it is clear that the Supreme Court’s holding in Crevier would limit Parliament’s ability to completely insulate the CITT from any Superior Court review. In the Customs Act, Parliament has not attempted to do so. Instead, it has provided an appeal mechanism, and simply limited what can be appealed to questions of law. Similar to the Code of Civil Procedure at the time Crevier was decided, the Federal Courts Actmakes clear that the traditional judicial review remedies provided for in that Act—injunction, certiorari, prohibition, etc.—are unavailable when a statutory appeal from an administrative decision is provided for: Federal Courts Act, s. 18.5.
In my view, Crevier supports the position that Parliament may restrict judicial review to questions of law. A statutory provision having this effect, such as section 68 of the Customs Act, meets any threshold established in Crevier.To hold otherwise would be to eliminate any possibility that Parliament could, via statute, restrict the ambit of judicial review of administrative action. What purpose would the specific provisions of the Customs Act, and many other federal statutes that restrict review, serve if recourse to the Courts could always be had on all issues under the general provisions of section 18 and section 28 of the Federal Courts Act? (at paras. 59-60)
Echoing Kristjanson J’s analysis, Near JA also commented that a process of concurrent appellate and judicial review proceedings “would be more burdensome and more complicated than the efficient and timely system of review contemplated by the Customs Act alone” (at para. 68).
But Gleason JA (LeBlanc JA concurring) took a different view on this issue. For her, at least some factual errors must be reviewable regardless of Parliament’s institutional design choices. Based on a magisterial review of the development of the “standard of review analysis” in Canadian law, culminating in Vavilov, she identified three fundamental propositions. First, the Supreme Court determined in Vavilov that “as a matter of principle, the availability of limited appellate review does not foreclose the availability of judicial review” (at para. 111). Second, there is no indication in Vavilov that privative clauses such as s. 67(3) of the Customs Act bar access to judicial review or curial oversight of any types of errors:
A complete bar on the availably of judicial review for any type of issue would offend the rule of law as the Supreme Court noted in Dunsmuir, a holding that was specifically endorsed in Vavilovat para. 24. Further, the Court in Dunsmuir and Vavilov did not overturn the previous decades-old case law determining that what were previously characterized as patently unreasonable factual errors, formerly called jurisdictional, remain reviewable, albeit now under the reasonableness standard (at para. 112. See also at paras. 82-87, discussing how “seriously erroneous factual determinations [can] constitut[e] patently unreasonable error”, and at para. 116).
Third, Vavilov expressly contemplates that “factual issues may give rise to unreasonable decisions” (at para. 113). As a result, a privative clause cannot be read “as barring access to judicial review for all factual issues” (at para. 116).
What weight, then, to accord to “institutional design”? Gleason JA explained that they are not necessarily meaningless:
Rather, they are part of the relevant statutory framework – an important contextual factor in determining the parameters of a reasonable decision according to Vavilov and the case law of this Court – and such clauses highlight the deferential nature of reasonableness review for decisions falling within the ambit of the clauses. I do not believe there is any other way to reconcile the collapsing of the patent unreasonableness and reasonableness standards of review into a single standard of reasonableness other than to recognize that review is available under the reasonableness standard for what were formerly characterized as patently unreasonable errors, which include serious factual errors, even in the face of a privative clause (at para. 117. See also Koebisch v. Rocky View (County), 2021 ABCA 265, at para. 24).
For Gleason JA, no tricky procedural issues would arise from concurrent appeals and judicial reviews: “Any overlap in proceedings could be addressed through joinder of an appeal with an application or other appropriate directions as might be required from time to time” (at para. 120). As I noted soon after Vavilov: over to you, rules committees.
Despite her different analytical approach, Gleason JA ended up concurring with Near JA as to the outcome of this particular appeal and her final landing point is perhaps not as far removed from Kristjanson J’s as one might initially think. This is because, both under Vavilov and under the jurisprudence of the federal courts, “the scope of review in respect of factual matters is limited, providing for intervention only in a narrow range of cases beyond those where there is a complete lack of evidence on a point” (at para. 121). Is this another way of saying that concurrent judicial reviews will only be entertained in exceptional circumstances where there is a limited statutory right of appeal? Their approaches are different — Gleason JA relies heavily on constitutional principle, whereas judicial discretion is fundamental for Kristjanson J — but the scope of judicial review is similarly circumscribed. Near JA, by contrast, proves implacable, with courts bound to respect Parliament’s institutional design choices, though in this particular case the theoretical disagreement made no practical difference.
This issue is going to run and run and run, as long the content of the core constitutional minimum of judicial review remains obscure.
One final issue is the standard of review for arbitration appeals, which continues to vacillate from province to province and even from court to court. I do not have anything to add to my last comment on the point, but this recent piece by Mancini and Plotkin is worth reading. Again, this one is set to continue running until there is an authoritative appellate consideration of the issue (ideally by the apex court).
This content has been updated on October 21, 2021 at 21:26.