Through a Glass, Darkly – Teaching Canada’s Administrative Law Standard of Review (Craig Forcese)

Craig Forcese is a Full Professor at the Faculty of Law, Common Law Section, at the University of Ottawa

The Supreme Court of Canada released its Dunsmuir decision in early March 2008, hours before my administrative law class at the University of Ottawa, Common Law Section, was to embark on study of standard of review in substantive administrative law.

The moment was bittersweet. It obliged me to set aside my carefully choreographed lecture notes on the Court’s pragmatic and functional test. I had, to that date, told the story of standard of review as a march toward abstraction. At the climax of that tale, I would recite, in tremolo-ing voice, portions of Justice LeBel’s dissenting j’accuse in Toronto (City) v. C.U.P.E.. I would include Barry J.’s observations in Miller v. Workers’ Compensation Commission (Nfld.):


In attempting to follow the court’s distinctions between “patently

unreasonable”, “reasonable” and “correct”, one feels at times as though one is

watching a juggler juggle three transparent objects. Depending on the way

the light falls, sometimes one thinks one can see the objects. Other times one

cannot and, indeed, wonders whether there are really three distinct objects

there at all.


By this point full of righteous indignation, I would end by reciting the parable of the Emperor’s new clothes, and the propensity of administrative law to perceive differences without replicable distinctions.  Or, depending on the year, I might suggest that application of the standard of review in Canadian administrative law had some of the hallmarks of the “know it when I see it” concept of obscenity, famously invoked by Justice Potter Stewart in the US Supreme Court.

To be clear, I did not build my teaching around this irreverent polemic. I did my best to square the doctrinal circle and teach the law of standard of review, as best as I could rationalize it. But my culminating classroom critique reflected, certainly, the vein of legal realism that colours my view of the law. Here, I sought to buck the propensity by law students to see Supreme Court decisions, not as precedent generated by the force of circumstances, but as a scripture, ordained and immutable. (Administrative law is a wonderful course to begin with Oliver Wendell Holmes’ observation that the live of the law is history, not logic.)

And I was also motivated by a more philosophical animus toward Canada’s standard of review; less a test producing predictable outcomes and more a conclusion reverse-engineered by the highest court. I posit that, given the scope of the regulatory state, no other of the law touches more people more directly than does administrative law. It should be the most accessible of disciplines. Instead, it had become among the most esoteric. That is deeply anti-democratic.

And so, reform of Canada’s standard of review was both overdue and welcome. In Dunsmuir, the Supreme Court did not adopt the formula I would have preferred. I am a distinct minority: I favour an abandonment of judicially-prescribed standards of review, predicated on an exegesis of variables invisible to the non-judicial eye. Let courts review on correctness, and if there is to be deference, let legislatures prescribe it, restoring true parliamentary supremacy in this area, creating predictability and making legislation speak plainly, not in code.

But Dunsmuir did, at least, abandon the meaningless tripartite standard of review. And it favoured a series of “default” indicators for the standard of review that at least removed the risk of multivariate considerations of uncertain breadth, each plausibly pointing in different directions and incapable of doing much more than wowing a citizenry with the splendor of the Emperor’s clothes.  Or perhaps not?  The frustrating paragraph 64 resuscitated the variables from the pragmatic and functional test, leading to endless classroom conversations not far-off Monty Python’s dead parrot skit (“it’s not quite pining for the field yet”).

More than that, the defaults invoked a new, uncertain vocabulary. What exactly was a question of law of central importance to the Canadian legal system, and outside the specialized expertise of the tribunal? And what was to be made of “true” questions of jurisdiction, distinguished from the fool’s gold of less-true questions of jurisdiction. Constitutional questions seemed straightforward, until Doré make them less straightforward.

But nevertheless, surely standard of review analysis now could be reduced to a few lines of judgment in a handful of cases, whose digestion would consume, perhaps, two class sessions? Not one-third of the term.  And that mattered – more experienced now, I was determined to recraft my manner of instructing administrative law. This discipline – the tool box for any citizen engaged with the state – should be approached as a Swiss army knife, not as an inquiry into the resident population of angels on pinheads.

And so, Dunsmuir promised simplification-reform contributed to my efforts to rebuild administrative law instruction around an active learning, client-centric model. My teaching now encourages students to solve the problems of hypothetical clients, whose problems are built-up in mimicked client files in our large-format class. If administrative law outcomes could rarely be predicated as an objective matter because of its oblique doctrine, it was better, I concluded, to teach it as an exercise in advocacy; ends-motivated analyses designed to serve the interests of clients. There is little point labouring over whether something should truly be reviewable on reasonableness or correctness, in response to some Platonic inner form. Rather, how can Dunsmuir be deployed to arrive at a result that best serves the immediate interests of a client’s case?

This has, I believe, been the right approach for first time learners of administrative law, and has facilitated both substantive and skills-based learning.  But it teeters on the edge, as the Supreme Court embarks on another round of baroque inquiry into standard of review. The rough-hewn, but still workable form of Dunsmuir is now encrusted with uncertainty: Doré, and review of constitutional “values” in discretionary decisions that aren’t quite as deserving of close judicial scrutiny as constitutional rights (or maybe just division of powers rules?) jeopardized in legislation; Newfoundland Nurses and the retroactive, forensic search for justifications sustaining doubtful decision-making; true errors of jurisdiction pooh-poohed as the unicorn of administrative law, and then unexpectedly discovered; and, most concerning, the shadow of Mordor: the excavation (maybe or maybe not?) of new “contextual indicators”, difficult to predict from Dunsmuir’s defaults and unmoored to even the oblique variables enumerated in the pragmatic and functional test.

Overreach in such cases is followed by subsequent retrenchment and clarification, reconsideration and minority views in new ascendance as the composition of the Court shifts: we are not far off a reversion to the pre-Dunsmuir situation. There is no plausible way a lawyer, let along their client, can predict a standard of review until definitive judgment is given by the Supreme Court, sometimes split, and sometimes also reversing lower court determinations that themselves split. Thirteen judges later, you have your answer.

Standard of review risks becoming law without a rule. Teaching standard of review is again the sort of undertaking whose frustrations are best communicated through emoji’s, not reasoned academic prose. And careful scientific study show this is not an idiosyncratic opinion: my irreverent tweets about recent developments in the standard of review caselaw are festooned with hearts, earning me scores of dedicated followers.

Treating standard of review as a subset of metaphysics seems certain way to cast the administration of justice in the administration of government into disrepute.  But fortunately, the Supreme Court seems to embark in a paradigm shift in this area every ten years or so. Whether what comes next will be better or worse remains to be seen. But the Court’s compass should be: predictive simplicity.

While we’re on that topic, is this a good place to talk about Baker and procedural fairness?

This content has been updated on February 26, 2018 at 20:37.