To my eye, Robertson J.A. lays out the basis for adopting an approach akin to the Chevron doctrine applied by U.S. courts. Where the law is clear, the courts must enforce it. Otherwise, any reasonable interpretation by an administrative decision-maker is permissible and should be upheld by a reviewing court. As Robertson J.A. puts it, "if the words are clear, they must generally prevail. If they are not clear, the words must yield to an interpretation that best meets the overriding purpose of the statute" (at para. 31). (I am not sure what significance to accord to the word "generally", the use of which is strange because nothing else in Robertson J.A.'s reasons suggests that he is setting out anything other than a rule that clarity will always prevail.)
Here are the key (in my view) points Robertson J.A. makes (at paras. 31-33):
The law has imposed a methodology, or legal framework, for addressing interpretative issues and it matters not whether the issue arises in the civil as opposed to administrative law context. In both instances, the reviewing court will be examining the methodology used to reach an interpretative result, including the question of whether the wording gives rise to “ambiguity”. As a general proposition, where deference is owed and there is ambiguity giving rise to two plausible interpretations, the one chosen by the adjudicative tribunal must be accorded deference. However, when it comes to the task of statutory interpretation, it is settled law that the words of an enactment must be read in their entire context and in their grammatical and ordinary sense, harmoniously with the scheme of the enactment, its object and the intention of the legislature or Parliament...
Hence, any reviewing court will look to see whether the interpretative process followed by the interpreter remains true to the legal principles which provide the legal framework and methodology for addressing interpretative issues...
Critical to the interpretative task is whether the text generates a real and not perceived ambiguity. But the provision in question cannot be read in isolation. Otherwise the plain meaning rule would defeat the requirement to conduct both a textual and contextual analysis...
Assuming the interpretative methodology leads to a finding of a genuine ambiguity and two plausible interpretations, the deference doctrine tells the reviewing court to accord deference to the one chosen by the adjudicative body.This approach is attractive. It has significant formal appeal: if the legislature has been clear, administrative decision-makers have no interpretive room, and the courts must give effect to the clear statutory language.
A related metaphor is that of the bell curve (proposed by amongst others Stephenson and Vermeule), with the lower edges of the curve identifying the clear limits and the hump the space within which the administrative decision-maker can choose between alternative interpretations.
Let me have a stab at explaining why I think this approach is wrong.
First of all, the proposition that "clarity" limits the authority of administrative decision-makers seems to me to be a reformulation of the proposition that courts ought to enforce jurisdictional limits (a proposition I quibble with). Exactly the same logic is employed by proponents of judicial intervention: clear limits have been established in legislation and should be enforced by courts. But the same problems recur: how can a "clear" provision be distinguished from an "unclear" one (especially once courts apply all the ample tools they possess to interpret statutory provisions); what basis is there to presume that a legislature granted authority to interpret only a subset of "unclear" provisions; and what about the relative institutional competence of the administrative decision-maker?
As I put it in "Deference on Questions of Law" (2011):
But even agreement that a question of law has been raised does not mean that the question of law has only one answer. Diligent, never mind ingenious, use of dictionaries and legislative history can conjure clarity where there appears to be none. Equally, as an American appellate judge has observed, extra-judicially, ‘virtually any phrase can be rendered ambiguous if a judge tries hard enough’. Even when following the narrow approach to judicial review of questions of law, the ‘search for meaning gives courts too much power to frustrate interpretative delegations when [the legislature] is likely to intend them and when they are likely to promote other normative values’. Indeed, ‘the question of clarity…is not binary; “clarity” and “ambiguity” are terms on a continuum that are meaningless without some implied threshold of certainty of meaning against which potential readings of the statute can be judged’. A reviewing court is impelled to make a prima facie interpretation of the statutory provision, which may colour its interpretative inquiry and lead it to insist upon clarity where in reality, none is present...
In fact, ‘there might be multiple valid interpretations of a statutory provision or answers to a legal dispute' and the legislature might have delegated a choice between different, valid interpretations of a statutory term to a delegated decision-maker. Such an aim could be frustrated by dogmatic judicial insistence that there is only one correct interpretation of a given statutory term. Adopting the [Chevron] approach as a default rule risks frustrating legislative intent, because that default rule may not reflect the intent of the legislature in a particular case...
The difficulty is that no yardstick is available to measure the accuracy of an allegedly clear answer. Put simply, just because a court defines ‘national security’ as ‘securing the freedom of individuals from terrorist attack’, it does not mean that the court is correct to do so. An ex post conclusion by a court does not settle the ex ante question of whether a clear answer exists. And for the purposes of the narrow approach, it is on the existence of a clear answer that the legitimacy of judicial intervention depends. If the ability of a reviewing court to intervene depends on knowing ex ante that a clear answer exists, the sphere in which a court can legitimately intervene is greatly reduced. (citations omitted)The gateway to deference is not the existence of ambiguity or vagueness, it is the initial delegation of interpretive authority.
Moreover, insisting on the application of the tools of judicial interpretation to determine whether statutory language is clear has the unfortunate effect of promoting a legalistic approach to administrative decision-making: if decision-makers want their interpretations to be upheld, they would be better to follow judicial approaches to statutory interpretation. Yet if the legislature had wished a legalistic approach to be taken to the interpretation of statutory provisions, they could have vested interpretation solely in the courts. To insist on the use of judicial tools is to undermine the legislative choice to empower administrative decision-makers.
It is also to discard the possibility of according "weight" to administrative interpretations of law, treating them as a serious factor which ought to influence the final interpretation arrived at by a court: this is an aspect of David Dyzenhaus' (much misunderstood) idea of "deference as respect". In the search for clarity, the considered conclusion of the administrative decision-maker is not a factor at all (it is heartening, then, that Robertson J.A. has elsewhere recognized the value of administrative interpretations of law).
How should a reviewing court proceed when determining the reasonableness of an administrative decision-maker's interpretation of law? At a high level of abstraction it is important to distinguish between interpretations that are "wrong" and those that are "unreasonable". The test for judicial intervention is "unreasonableness", not "wrongness". It is not because the decision-maker fails to live up to some judicially imposed ideal that judicial intervention is justified. Something more is required. Otherwise, judges would be intervening because the administrative decision-maker was "wrong", an approach entirely at odds with administrative autonomy.
What is required is for the applicant to demonstrate some error on the part of the administrative decision-maker. Some examples may be helpful. A prima facie departure from statutory text requires an explanation. So does ignoring a key term altogether (the very problem in Small: see paras. 37 and 39). An interpretation which runs contrary to the purpose of the statute would also be problematic.
If the decision-maker has not cogently explained such interpretive errors, judicial intervention is justifiable. This approach is consistent with the approach taken to judicial review generally. Quite properly, the onus is on the applicant to demonstrate reviewable error; it is not for the administrative decision-maker to demonstrate that it has acted in a manner that is beyond reproach. Judicial intervention on the basis of perceived clarity, or a failure to follow a prescribed approach to statutory interpretation, would compromise the autonomy of administrative decision-makers. Robertson J.A.'s approach, though attractive, ought to be rejected.