In this post, I will assess some recent appellate pronouncements on the relationship between deference and reasonableness. I find these pronouncements somewhat troubling and see them as indicating that Canadian courts will move away from their unique approach to review of questions of law and towards something like the Chevron deference applied by American courts. I should also say that I will be developing these thoughts at greater length in two projects in the near future: the chapter on "Légalité" for the Jurisclasseur -- Droit Administratif; and a chapter in Mapping Taggart's Rainbow, a collection edited by Mark Elliott and Hanna Wilberg.
My concern is that courts are paying lip service to deference by applying the reasonableness test in a non-deferential manner, indeed, in such a way as to undercut deference.
In an important and intelligent set of reasons, Stratas J.A. has, in my view, accurately laid out the current Canadian position: Canada (Attorney General) v. Canadian Human Rights Commission, 2013 FCA 75. (Evans J.A., in an extra-judicial essay in the most recent issue of the Canadian Journal of Administrative Law and Practice, has echoed the comments of his colleague; and before his retirement from the Supreme Court, Binnie J. recently expressed similar sentiments in a concurring set of reasons.)
The issue here was whether the Canadian Human Rights Tribunal had applied the correct test in holding a claim of discrimination inadmissible. The claim was in relation to the provision of services to First Nations. As Stratas J.A. explained, the Tribunal dismissed the claim for the following reason: "there is no relevant comparator group because the Government of Canada does not provide welfare funding for any other children" (at para. 7). Stratas J.A. agreed with the Federal Court that this conclusion was unreasonable, based on the text, context and purpose of the Act, Canada's international obligations, the equality jurisprudence of the Federal Court, and the reduced importance accorded by the Supreme Court of Canada to the need to demonstrate comparator groups.
Of greater importance for present purposes are these comments:
 In this case, the range [of acceptable and defensible outcomes] is relatively narrow. The Tribunal’s decision primarily involves statutory interpretation – a matter constrained by the text, context and purpose of the statute. It also involves equality law – a matter constrained by judicial pronouncements. In this case, the Tribunal had less room to manoeuvre than in a case turning upon one or more of factual appreciation, fact-based discretions, administrative policies, or specialized experience and expertise not shared by the reviewing court on the particular point in issue.
I criticize Mowat at some length in this article, but let me push back further against Stratas J.A.'s comments. The Supreme Court’s decision in Mowat, supra – also involving a review of the Tribunal’s interpretation of the Act – illustrates this well. There, the Supreme Court reviewed the Tribunal on the basis of the deferential standard of reasonableness. However, acting under that standard, the Supreme Court engaged in an exacting review of the Tribunal’s decision, a review more exacting than that of the Federal Court in this case. Some might describe what the Supreme Court did in Mowat as disguised correctness review. I disagree. Mowat is reasonableness review, still deferential, conducted in recognition that, as far as the Supreme Court was concerned, the Tribunal had only a narrow range of acceptability and defensibility open to it, given the constrained nature of the matter before it. Within that range, the Tribunal was entitled to deference.
First, I am not sure I agree with the premise that interpretations of law have a necessarily narrower range than exercises of discretion. For one thing, as a practical matter, when a question as to the exercise of discretion comes before a reviewing court, it will often be posed in a binary manner: should the decision. -maker have exercised its discretion in this way, or should it not have? But when a question of law is at issue, multiple interpretations will be open to the Court. (See also Justice Breyer's short, but very interesting, opinion from this week in Wos v. E.M.A.)
More generally, as a theoretical matter, the premise does not hold in all cases. Take Mowat, for example, where the question was the meaning of "costs" in the Tribunal's governing statute. Here, "costs" could mean a whole host of things: all disbursements, all legal disbursements, legal costs on a partial, substantial or full indemnity scale, or maybe even damages. And in exercising a discretion to award costs, the Tribunal is somewhat constrained, primarily by the claims of the parties. It cannot just pick a random number somewhere between the two positions. There must be some reasoned explanation for its exercise of discretion. In short, the exercise of discretion as to awarding costs could be more constrained than the interpretation of the term "costs".
Second, as will be clear from the previous paragraph, I do not agree with the premise that a clear distinction can be made between "law" and "discretion" (see pages 16 and 17 of this article for a brief and over-simplified explanation). I think this premise supports Stratas J.A.'s explanation, but I do not think the premise is tenable. As the Mowat example helps to demonstrate, "law" and "discretion" are very similar creatures. Any reasoned explanation of an exercise of discretion must be made by reference to the statute; and any interpretation of a statutory provision, inasmuch as it presents a choice to the decision-maker, involves an exercise of discretion. Without the ability to make a clear distinction between "law" and "discretion", I doubt the appropriateness of saying that a decision-maker is more constrained in interpretations of "law" than in exercises of "discretion".
Third, Stratas J.A.'s approach requires reviewing courts to determine the boundaries of the "range of acceptable outcomes". The danger here -- of which I think Mowat is an excellent example -- is that in order to determine the "range", the reviewing court has to determine what the best interpretation of the statute is. But as soon as one opens the door to the concept of "range", one opens the door to judges imposing their preferred interpretation of the statute. This even though the Supreme Court has said that administrative decision-makers' interpretations of their governing statute are entitled to deference. This mode of application of the reasonableness standard, then, contradicts the presumption of deference.
Now, this is not a new problem: much the same criticism has been made of reviewing courts' treatment of exercises of discretion. Ironically, Stratas J.A. himself has eloquently explained the point:
There is no contradiction here in Stratas J.A.'s thinking (and as I said, he is expressing the conventional judicial wisdom), but this passage underpins the importance for him of a distinction between law and discretion. Under the Supreme Court’s approach, we do not determine what considerations are relevant and then impose our determinations of relevance on the tribunal. Rather, the tribunal is given “substantial leeway…in determining the…‘relevant considerations’ involved in a given determination,” and then we engage in reasonable review of what the tribunal has done: Baker, supra at paragraph 56. Reasonableness review is supposed to be truly deferential review: Dunsmuir, supra at paragraph 47.
As I suggested at the outset, we are not very far here from Chevron deference. A recent decision of the Ontario Court of Appeal exemplifies this.
The question in Ontario (Alcohol and Gaming Commission of Ontario) v. 751809 Ontario Inc. (Famous Flesh Gordon's), 2013 ONCA 157 was whether the Board of the Commission should find that an individual's membership of the Hell's Angels, a criminal organization, justified revoking a liquor licence. The Board thought not, but the Ontario Court of Appeal thought differently: in its view, the legislation "requires that past and present conduct not be confined to the individual’s operation of the licensed establishment" (at para. 29). Goudge J.A. went on to state:
The references to "plain language" and "clear purpose" have a very American ring. Under Chevron deference, courts are to uphold reasonable interpretations of statutory provisions, as long as the statute is unclear. The problem with this formulation is that, with even a minimal amount of judicial imagination, clarity can always be conjured up. I conclude that the Board erred in law in its interpretation of s. 6(2)(d) of the Act. It may be that if this case is seen as an interpretation by the Board of its home statute, it may be owed some deference in this court. That need not be decided here because, even if that is so, my view is that the Board’s erroneous interpretation of the test does not fall within a range of possible acceptable interpretations, given the plain language of the section and the clear purpose of the legislation. It is not a reasonable interpretation of the legislative test. It is outside any margin of appreciation that deference requires. (my emphasis)
Moreover, in the Famous Flesh Gordon's scenario, rather than starting with the presumption that administrative interpretations of law should be displaced only if they are demonstrably unreasonable, Goudge J.A. started by adopting his preferred interpretation of the law and then measuring the Board's interpretation against it. This temptation is ever-present with Chevron deference.