In Dunsmuir v New Brunswick, 2008 SCC 9,  1 SCR 190, the Supreme Court sought to bring clarity and (relative) simplicity to the law of judicial review of administrative decisions, which as it acknowledged with some understatement “ha[d] not been without practical and theoretical difficulties, [or] free of criticism”.  But subsequent decisions, digesting, developing, and departing from Dunsmuir, revealed the futility of its promise of simplifying the law of judicial review in Canada. Practical and theoretical difficulties, and criticism, still abound ― at least when it comes to review of administrative decisions on questions of law. The law on this issue, as other contributions to this symposium note, is in a parlous state. The Supreme Court’s decisions fail to provide guidance to litigants and to lower courts. They are difficult to understand, unrealistic, and appear to do something very different from what they say they do. A fundamental re-assessment, of a magnitude at least equal to that of Dunsmuir, seems to be inevitable.
In my view, one reason why Dunsmuir failed to simplify and clarify the law of judicial review once and for all is the weakness and incoherence of the justifications it provided for judicial deference to administrative decisions. In this post, I review these justifications and argue that none of them can account for the broad scope of deference that Dunsmuir and subsequent cases mandate. Any attempt to reformulate the law of judicial review in the future must acknowledge the weakness of the available explanations for deference, and can only require courts to defer to administrative decision-makers in narrow circumstances where such deference would be well and truly justified.
* * *
Dunsmuir proclaimed that “determining the applicable standard of review is accomplished by establishing legislative intent”.  If courts were sometimes, or often, to defer to administrative decision-makers’ interpretations of law, that was because legislatures wanted them to do so. “The existence of a privative or preclusive clause”,  providing that an administrative decision was not to be interfered with by the judiciary, was the indication par excellence of the enacting legislature’s desire to commit the determination of issues arising out of the operation of a statutory scheme to a tribunal rather than a court.
But Dunsmuir itself and subsequent cases undermined what the simplicity of a hypothetical regime where privative clauses trigger deference ― and their absence, logically, leads to non-deferential review. For one thing, as Dunsmuir acknowledged, to the extent that they purport to oust judicial review completely, privative clauses cannot be given their full effect, because under the constitution “neither Parliament nor any legislature can completely remove the courts’ power to review the actions and decisions of administrative bodies”.  But even putting this constitutional limitation to one side, under the Dunsmuir framework a privative clause is not the only signal of deference. Indeed, according to Edmonton (City) v Edmonton East (Capilano) Shopping Centres Ltd, 2016 SCC 47,  2 SCR 293, even a provision creating a right of appeal on a question of law, which seems like an explicit indication that a legislature does not want courts to defer to administrative decisions, is not enough to oust a “presumption of deference”, which thus takes on a life of its own, unmoored from legislative intention.
Another reason for judicial deference to administrative decisions, according to Dunsmuir, is that “certain questions that come before administrative tribunals” ― including questions of law ― “do not lend themselves to one specific, particular result”.  The multiplicity of possible answers to the questions facing administrative decision-makers combined with the need to “respect … the legislative choices to leave some matters”  to their arbitrament to justify deference.
The problem here is that there is little reason to think that the sets of questions of law that “do not lend themselves to one specific particular results” and of questions on which deference is required under the Dunsmuir framework (or the Dunsmuir framework as modified by subsequent Supreme Court decisions) are identical. Dunsmuir called for deference “where a tribunal is interpreting its own statute or statutes closely connected to its function, with which it will have particular familiarity”  or related common law rules. Yet the great variety of statutes setting up administrative tribunals, and indeed of particular provisions within any one of these statutes, makes it unlikely that all of the interpretive questions to which they give rise lack definitive answers. Perhaps the suggestion is that the very legislative choice of setting up administrative tribunals to address these questions means that legislatures think that these questions lack definitive answers, but that too seems implausible. A legislature may wish to set up an administrative tribunal for any number of reasons that have nothing to do with the existence or not of clear answers to interpretive questions: cost-effectiveness, accessibility, the need to execute a law (and not just adjudicate disputes), even patronage.
Finally, Dunsmuir insisted that courts need to defer to administrative decision-makers out of “respect … for the processes and determinations that draw on particular expertise and experiences”.  Since then, as the dissenting opinion in Edmonton East noted, “the notion of ‘expertise’ has become a catch-all trigger for deferential review”,  even though “this presumption of expertise has rarely been given much explanation or content in our jurisprudence”.  Meanwhile, the majority opinion suggested that expertise was independent of the qualifications or functions of administrative decision-makers. The very existence of a specialized tribunal made it an expert.
Perhaps the popularity of the presumption of expertise is due to the Supreme Court’s realization that the other justifications for deference are unpersuasive. Perhaps it is a convenient means to disclaim responsibility for decisions taken elsewhere (similarly, for instance, to the Supreme Court’s insistence, starting in Canada (Attorney General) v Bedford, 2013 SCC 72,  3 SCR 1101, on deference to first-instance judges’ findings of legislative fact). Other contributors to this symposium challenge the notion of administrative expertise as a foundation for deference in greater detail. Suffice it for me to say that, just like legislative intent and the impossibility of a definitive answer, expertise is often a judicial fiction and thus an unconvincing justification for deference.
All this is not to say that there are no cases in which deferential review on a reasonableness standard would not have been intended by a legislature setting up an administrative tribunal, or in which there is indeed a multiplicity or a range of plausible answers to legal questions, or in which the administrative tribunal is (more) expert (than the reviewing court). My point, rather, is that, at a minimum, these justifications do not support deference across the board when administrative decision-makers are interpreting their “home statutes”, which Dunsmuir, and especially cases like Edmonton East require. Indeed, these justifications can operate at cross-purposes, as when the legislature authorizes appeals on questions of law from the decisions of expert tribunals, or when administrative decisions are insulated by privative clauses from review of their answers to questions which do in fact “lend themselves to a specific result”. The Supreme Court’s jurisprudence suggests that all these conflicts must be resolved in favour of deference, but it does not provide any explanation for why this is the case.
* * *
The Supreme Court, presumably, can see these difficulties as plainly as its critics. They might account for at least some of the frequency with which ostensibly deferential review exhibits no sign of deference at all ― another issue that other contributions to this symposium highlight. The only way forward, in my view, is for the Court to allow its explanations to match its decisions, and abandon the pursuit, or the pretense, of across-the-board deference. If deference is ever appropriate, it can only be justified with reference to the circumstances of particular cases. This is not the place for a full exploration of the circumstances in which courts ought to defer to administrative decision-makers on questions of law, if indeed there be any. I will, however, venture a few observations.
First, it should not be presumed that deference is due to all the decisions of a particular administrative tribunal or type of tribunal, or to tribunals interpreting a type of statute. As the dissent in Edmonton East pointed out, a single tribunal may be called upon to decide different types of issues, some of which implicate its expertise or involve policy considerations, while others do not. In particular, even if deference is appropriate to tribunal decisions fleshing out the meaning of vague terms such as “just”, or “reasonable”, or “in the public interest”, it does not follow that it would be called for when the same tribunals answer questions about, say, the relationship between provisions of a statute (even their “home” statute) or two related statutes. The former type of question calls upon the tribunals’ specialized knowledge of the standards of justice or reasonableness, or the requirements of the public interest in the field they regulate. The latter is concerned with more general legal skills which courts possess no less, and often more, than tribunals.
Further, even when dealing with a type of question answers to which would normally attract deference, administrative decision-makers are sometimes so constrained by judicial precedent that their decision-making is not meaningfully different from that of a trial court striving to apply appellate authority. The decision that led to in Quebec (Commission des normes, de l’équité, de la santé et de la sécurité du travail) v Caron, 2018 SCC 3 is a good example, as the relevant analysis concerned the scope of the jurisprudence of the Québec Court of Appeal. In similar circumstances, a trial court’s conclusions would not be entitled to any deference on appeal. Why should the administrative decision-maker’s? Assertions to the effect that judicial review and appellate review are not identical are not convincing in those cases where the nature of the decision under review is such that they practically are.
Finally, and more substantively, it is important to recall what is at stake in judicial review of administrative decisions. Proponents of deference often think of it as a means of protecting the decisions of an administrative state devoted to economic regulation in the name of social justice, or at least of enlightened technocracy. But there is much more to the administrative state economic than labour boards or arbitrators, whose decisions supply a disproportionate share of material for the Supreme Court’s administrative law decisions. The law of judicial review of administrative action applies also to the review of correctional authorities, professional licensing bodies, immigration officers, human rights tribunals, even universities and municipalities, and much else besides. People’s ability to enjoy their property or to practice their profession, their right to enter into or to remain in Canada, even their liberty (or at least the conditions of their detention, which the Supreme Court recognizes as a liberty interest) can depend on the way in which an official or a body exercising powers (purportedly) delegated by a legislature interpret the law. Is it enough to tell them, as Dunsmuir effectively does, that it is sufficient that the interpretation that causes them to lose these rights or benefits be justified, transparent, and intelligible?
* * *
Dunsmuir sought to simplify Canadian administrative law by setting out a unified analytical framework based on “the structure and characteristics of the system of judicial review as a whole”.  As part of this process, it set out a number of justifications that were supposed to support a wide-ranging policy of judicial deference to administrative interpretations of law. Well-intentioned as it was, the attempt did not succeed. The justifications advanced in Dunsmuir cannot justify deference in many cases where the Supreme Court said it is due. The weakness of, and occasional conflicts among, the justifications for deference advanced in Dunsmuir have fostered renewed confusion in the law of judicial review.
It is important that this confusion be eliminated. Canadian administrative law is reaching a point where it can scarcely be called law at all, such is its inability to provide guidance to those who must apply it or predict how it will be applied. Yet ― perhaps paradoxically ― the way to clarity passes not through the application of a single all-encompassing principle, but through greater attention to the circumstances of individual cases. As these circumstances vary, so must the applicable rules. It is Dunsmuir’s attempt to deny or at least avoid this complexity that is responsible for its failure.