A Decade of Dunsmuir: Please No More (Hon. David W Stratas)

David Stratas is a judge of the Federal Court of Appeal


The views expressed in this post are the author’s own and are offered for education and discussion purposes only.

Dunsmuir is law. Like all Supreme Court decisions, it binds me. Nothing in this article suggests otherwise. If you come to my court and tell me that I must disregard something in Dunsmuir and if you do not offer a Supreme Court authority that allows me to do that, I am duty-bound by law to reject what you say.

But I do hope that the Supreme Court will revisit Dunsmuir and the cases that purport to follow it. Judging by the articles in the Dunsmuir Decade series posted over the last month, an overwhelming number of leading Canadian commentators find much of the jurisprudence confused, incoherent and troubling.

This is not to besmirch in any way Dunmsuir’s dedicated, accomplished authors, the strong court that concurred with them, or the talented jurists that decided the cases after Dunsmuir.

Rather, the flaw in Dunsmuir and its progeny is the same one that has undermined this area of law in Canada at all levels of court for decades: the tendency of individual judges to decree operational rules based on their individual views rather than on fundamental, settled doctrine and well-accepted principles.

Dunsmuir and its progeny are like a great skyscraper—but one built on foundations of sand, not bedrock. Thus, it is no surprise that questions of inconsistency and incoherence concerning Dunsmuir and its progeny swirl about—and as the years go by, with increasing ferocity.

This article tries to make the case for why the Supreme Court should revisit Dunsmuir.

In another contribution to the Dunsmuir Decade series, I try to paint a picture of a happier future, one without Dunsmuir, based not on the sand of the shifting say-so of individual judges, but rather one anchored in the bedrock of settled doctrine and well-accepted principles.

Dunsmuir and the law of substantive review

Ten years ago, Dunsmuir was born. But Dunsmuir fits within a troubling context.

Questions concerning the standard of review for the substance of administrative decisions or, as I call it, the law of substantive review have vexed Canadian law for four decades now: see D. Stratas, “The Canadian Law of Judicial Review: A Plea for Doctrinal Coherence and Consistency” (2016), 42 Queen’s L.J. 27 at p. 29.

Dunsmuir was supposed to settle these questions once and for all. Specifically, it was to “develop a principled framework that is more coherent and workable” (Dunsmuir at paras.1 and 32).

But ten years after Dunsmuir these questions continue to roil and fester. In fact, this area of law has never been more confused. How come?

The nature of administrative law as the cause for confusion?

Many might blame the nature of administrative law itself.

No less an authority than the renowned, former Chief Justice, Beverley McLachlin, recently described administrative law as “a barbed and occluded thicket” where we find “only confusion”: “Administrative Law is Not for Sissies: Finding a Path through the Thicket” (2016) 29 Can. J. Admin. L. & Prac. 127.

The highly regarded Chief Justice of New Zealand, Sian Elias, recently described administrative law as “untidy and tentative”: Sian Elias, ‘The Unity of Public Law?’ in M. Elliott, J. Varuhas and S. Wilson Stark, The Unity of Public Law? Doctrinal, Theoretical and Comparative Perspectives (forthcoming, Hart Publishing, 2018).

I have called the law of substantive review a “never-ending construction site where one crew builds structures and then a later crew tears them down to build anew, seemingly without an overall plan”: Stratas, above at p. 29.

But these are just symptoms stemming from a deeper source.

The real cause of the confusion

In this area of law, individual judges at all levels of court—for the best of motives—have been decreeing rules and outcomes based on their own personal notions of the proper role of the judicial branch vis-à-vis administrative decision-makers. Judges start with what they personally think is best, rather than deducing the rules and outcomes from well-settled and fundamental doctrine and principles.

I think this is evident from the reasoning in the cases themselves. The analysis below makes this case. And certainly many in this online series, the Dunsmuir Decade, have pointed to a lack of principled application of Dunsmuir in later cases.

In my experience at judicial education conferences on administrative law, many judges are eager to volunteer their own personal opinions, inclinations and say-so’s about the separation of powers and the proper role of the courts—e.g., “I think judges should readily interfere with the decisions of non-judges,” “I think that we have to control tribunals,” “I think judges should respect the considered opinions of expert adjudicators,” “I don’t think people on these boards have expertise,” “I think as judges we know what’s best,” or “I think judges should do what’s fair.”

On these things, everyone seems to have a personal view held in good faith. Absent constraints supplied by settled doctrine and well-accepted principles in this area, views such as these inexorably find their way into judgments.

To some extent, academics are oriented in this way too. Many, including some contributors to the Dunsmuir Decade series, seem inclined to fashion operational rules to advance causes or outcomes they personally consider important, rather than basing them on settled doctrine and well-accepted principles.

Doctrine and principle: why they matter

Sourcing or deducing operational rules from settled doctrine and well-accepted principle enhances their acceptability and permanence. Judges are more likely to follow authority when it is rooted in more than just the say-so, personal opinions or political objectives of an earlier judge.

There’s a good reason for this.

It’s one thing to depart from operational rules in an earlier authority that are unsupported by settled doctrine and well-accepted principle; it’s quite another to depart from an authority supported by settled doctrine and well-accepted principle.

To disregard the former is to disobey a retired judge’s personal views and say-so. That’s something quite easy to do. Even if one tries to obey the earlier authority, the lack of underlying doctrine or principle underneath it means it is quite easy to go astray.

To disregard the latter is to throw settled doctrine and well-accepted principle—the collected wisdom of many judges tested in the crucible of many decided cases over a long time—into the waste bin. That’s something quite hard to do. And the presence of underlying doctrine or principle guiding the way furthers consistency and coherence in the later jurisprudence.

Judicial interference with the decisions of administrators can be a controversial thing. The sting of controversy can be reduced when judicial interference is prompted by settled doctrine or the responsible, incremental extension of legal doctrine achieved through pathways of objective legal reasoning practised by judges for decades, if not centuries—not judges’ whims and idle thoughts, their own world views, their own freestanding opinions of the moment about what is just appropriate and right or their own personal sense of what in the circumstances is fair and nice: see the authorities at nn. 135-140 of “The Canadian Law of Judicial Review: A Plea for Doctrinal Coherence and Consistency”, above. For more on this, see the speech here and the recent presentation here.

Finally, under the constitutional principle of the rule of law, litigants deserve equal application of law and equal treatment under that law: Imperial Tobacco. To the extent possible, judges should be giving similar rulings to similarly situated people. The only way that can be done is through reliance on ideas and concepts binding upon them, a body of settled doctrine and well-accepted principle—not just the say-so, instinct or whim of some judge that “something has gone wrong”: see generally, AI (Somalia) v Immigration and Protection Tribunal, [2016] NZHC 2227 at paras. 2 and 38-44, commenting on R. v. Panel on Take-overs and Mergers, Ex parte Guinness Plc, [1990] 1 Q.B. 146 (Eng. C.A.) at p. 160.

In the law of substantive review, is doctrinal decision-making possible?

But here we hit a roadblock. Chief Justice Elias of New Zealand has queried whether in this area of law “the search for better doctrine is ultimately doomed”: Elias, above.

On this, count me with Chief Justice McLachlin. On Chief Justice Elias’ query, she would demur. In her view, “if the rule of law is to prevail, a way [through the ‘barbed and occluded thicket’] must be found”: McLachlin, above. In other words, Chief Justice Elias’ question misses the point: we must embark on the quest for doctrine because the rule of law requires no less.

In my other contribution to the Dunsmuir Decade series, I demonstrate that settled doctrine and well-accepted principles do exist. And from these, we can deduce many operational rules governing the law of substantive review. But first, let’s turn specifically to Dunsmuir.

What Dunsmuir has wrought

Dunsmuir began well. At the start of its analysis, it set out (at paras. 27-31) two settled, well-accepted constitutional principles that inform this area of law: legislative supremacy and the rule of law.

One might have thought that from these two principles, certain operational rules governing how we go about applying the law of substantive law would be deduced and presented.

But no, Dunsmuir did not do that. Instead, Dunsmuir proceeded to ignore the two principles completely. It simply decreed certain operational rules to which the then-judges on the Supreme Court could agree. For reasons I shall explain, the operational rules in Dunsmuir seem to be a compromise of the moment rather than something founded on settled doctrine and well-accepted principles.

Dunsmuir set out two markedly different operational rules for determining the standard of review: applying certain prima facie rules, later described as presumptions, and assessment of a list of contextual factors (at paras. 53-61). Why? Who knows? Dunsmuir simply decrees these as the two operational rules.

For the most part, the presumptions (at paras. 53-54 and 56-61) are decreed out of thin air. They are not rooted in settled doctrine and well-accepted principles. Nor is the use of contextual factors (at para. 55) explained or justified.

Certain questions lay bare the vacuum of doctrine and principles. Why have presumptions? Why these particular presumptions? In particular, why should there be a general rule of deference for administrative interpretations of home statutes? What is the effect of the presumptions? When do we cast aside the presumptions and consider the contextual factors? Why aren’t the presumptions just more contextual factors to consider? How do we consider the contextual factors? Why these particular contextual factors—ones substantially the same as before—when Dunsmuir’s aim was to reform the law and proceed in a different direction (Dunsmuir at paras.1 and 32)?

Dunsmuir and, more particularly, its progeny decree other things too.

Reasons need only show the administrator was alive to the issue: Newfoundland Nurses. Why such a low standard? Who knows? Others’ attempts to formulate jurisprudence based on doctrine and principle went unassessed and unmentioned.

In many cases, legislators have made administrators the merits-deciders over a subject-matter, restricting (at best) reviewing courts to a reviewing function. But Dunsmuir at para. 48 allows reviewing courts to delve into the merits and supplement administrators’ reasons on the merits to uphold the outcomes they reach: Dunsmuir at para. 48. Why? At the level of doctrine and principle, this makes no sense: see Lemus [Fed. C.A.] at paras. 27-38. The justification offered in para. 48 of Dunsmuir? A single, unexplained and unsupported sentence plucked out of context from an academic article that has little to do with the mechanics of judicial review.

In appeals from decisions of reviewing courts, the decisions are subject to de novo review: Agraira at paras. 45-47. Why? In every other appellate context, unless the first instance judge makes an error of law or of extricable legal principle, the standard is palpable and overriding error: Housen. Why a different approach in this area of law? Given the importance of access to justice and minimization of expense (see Hryniak), why require the appellate court to conduct a complete rerun of the exercise?

Perhaps the most distressing part of Dunsmuir is its failure to say anything principled or doctrinal about the deferential standard of reasonableness, a new standard it created from the former patent unreasonableness and reasonableness standards. Dunsmuir decreed a few words like “acceptability,” “defensibility,” “transparency,” “intelligibility,” and “justification”: Dunsmuir, para. 47. But, by themselves, these are all just nice-sounding words in the eye of the beholder, not objective legal standards capable of being applied judicially.

As a result, the reasonableness standard is becoming a standardless sweep that allows judges at all levels to do whatever they subjectively feel is best—something that tears at the notion of the rule of law: see the discussion in AI (Somalia), above. All too often these days, some simply quote the nice-sounding words in paragraph 47 of Dunsmuir, offer the meaningless bromide that “reasonableness takes its colour from the context” without explaining the context, and then conclude that the decision is reasonable or unreasonable without much or any explanation at all.

The result? Empty words that serve as a not-so-transparent fig leaf for a purely personal, impressionistic, idiosyncratic thumb’s up or thumb’s down on an important administrative decision. If this area of law is to enjoy legitimacy and some degree of permanence, it must strive for better than that.

Other post-Dunsmuir cases

Without principle or doctrine supporting the decrees of Dunsmuir and succeeding cases, it’s all too easy for their holdings to be regarded as just the say-so of a certain group of judges, most of whom are now retired. Thus, it is no surprise that we see departures from them in later cases. Here are just a few examples.

– I –

The presumption of reasonableness review for administrative interpretations of home statutes. Significant minorities of the Supreme Court have repeatedly doubted this and re-questioned it. Stare decisis be damned! See, e.g., the minority holdings in Alberta Teachers, Wilson, Edmonton East, and Guérin, to name a few.

Most recently, attempts have been made to supply judicial policies in favour of the presumption but many will find them unpersuasive; they have not yet convinced almost half the Supreme Court: see Edmonton East at paras. 22, 33 and 35. See also the critique in D. Stratas, The Canadian Law of Judicial Review: Some Doctrine and Cases (October 21, 2017) at pp. 76-77.

For one thing, the majority in Edmonton East suggests that their judicially constructed presumption concerning administrative interpretations applies notwithstanding any contrary indicators in legislation; it is for the legislator to go further and specifically oust the judicially constructed presumption. But the court must respect legislative intention in the first place, not ignore it. Since when do judicial rules oust legislation? See ibid. at pp. 9-14. And inexplicably, in earlier cases, the Supreme Court has been attentive to contrary indicators in legislation: e.g., Tervita, Rogers.

– II –

The nature of reasonableness. Dunsmuir told us (at para. 47) that reasonableness is a deferential standard. But in tens of cases the Supreme Court’s application of the reasonableness standard is indistinguishable from correctness review: see the Dunsmuir Decade articles by D. Mullan, M. Liston; see also Stratas, Some Doctrine and Cases, above at pp. 52-53.

And reasonableness is said to be a single standard of deference that does not vary with the context: see, e.g., Khosa; Wilson at para. 18. But often it is said to “[take] its colour from the context” or depend on “all relevant factors”: see, e.g., Wilson at paras. 22 and 73; Catalyst Paper Corp. at para. 18; Irving Pulp & Paper, Ltd. at para. 74; Halifax at para. 44.

Discussion of margins of appreciation or ranges of outcomes that vary depending on the context are said by some, without explanation, to be “not an appropriate development”: see., e.g., Wilson at para. 73. But others enthusiastically embrace the idea: see Catalyst Paper Corp. and McLean; and see, most recently, Williams Lake at para. 34; and see many, many more cases where the Supreme Court evidently grants certain administrators more leeway or a wider margin of appreciation than others.

– III –

The presumption of correctness review for “true questions of jurisdiction” (Dunsmuir at para. 59). Without a doctrinal or principled justification for the presumption, Dunsmuir could not define it with precision. The best Dunsmuir could do was to cite United Taxi, a vires case. And what is a true issue as opposed to a not-so-true issue?

Soon after Dunsmuir, the Supreme Court has repeatedly queried whether “true questions of jurisdiction” exist (see e.g. Alberta Teachers at para. 34; Edmonton East at para. 26). Most firmly, it unanimously rejected the idea of preliminary or key questions that go to jurisdiction in Halifax. Yet, within a couple of years of that, a minority of the Court said that questions of this sort do go to jurisdiction and are a central, defining feature of judicial review: Guérin at paras. 66-68.

– IV –

Correctness for questions of central importance to the legal system. All we know from Dunsmuir (at paras. 53-54 and 60) is that these are “questions [that] require uniform and consistent answers” This has never been defined in any detail and no doctrinal basis has been supplied.

Left undefined and with no expressed doctrinal basis, “central importance” is in the eye of the beholder. So we see randomness. Time and time again, the Supreme Court has declined to apply this presumption even though it has granted leave on the ground of public importance: see, e.g., Nor-Man; Mowat; McLean; Irving; Wilson and many more. But occasionally it does apply this presumption: see, e.g., Saguenay; University of Calgary.

Without a doctrinal basis, this presumption will be used by some as a fire alarm that can be pulled whenever they choose, for whatever reason, in order to evacuate themselves from reasonableness review.

– V –

Constitutional issues. While constitutional questions are to be reviewed for correctness (Dunsmuir at para. 58), administrative decisions based on a consideration of constitutional values, even those where constitutional considerations predominate, are to be reviewed for reasonableness (Doré).

But significant numbers of judges on the Supreme Court have turned their backs to this approach (see, e.g., Loyola), again without explanation—as have some other judges on lower courts (see E.T. [Ont. C.A.] at paras. 108-125 and see the extra-judicial lecture here).

And Doré seems inconsistent with settled doctrine and some longstanding, fundamental and well-accepted principles: see D. Stratas, “The Canadian Law of Judicial Review: A Plea for Doctrinal Coherence and Consistency” (2016), 42 Queen’s L.J. 27 at pp. 34-35 and 37-38.

– VI –

Supplementing administrators’ reasons (Dunsmuir at para. 48). Back in 2013, two decisions of the Supreme Court—released just one day apart—conflicted on just how far judges can supplement: Newfoundland Nurses and Alberta Teachers, discussed in Lemus, above at paras. 27-34. And more recently, two decisions of the Supreme Court just two weeks apart also seem to conflict. Indeed, the author of the first decision (Delta), Chief Justice McLachlin, joined the full-throated dissent by Justice Brown on this point in the second decision (Williams Lake).

As former Justice Evans has noted, this jurisprudence incentivizes administrators to write reasons that “need merely hint at the basis of the interpretation” and leave it to “be elaborated and patched up [by reviewing courts] on judicial review.” This is contrary to law: the legislator has given the administrator the job of deciding and explaining itself so it should do just that!

In addition, para. 48 of Dunsmuir now would seem to allow us to review decisions that were never made and fashion reasons for them: see Edmonton East, a doctrinally curious case. It also goes against the principled idea in Alberta Teachers and Okwuobi that, as a general rule, issues that should have been raised before the administrator should not be determined on judicial review.

Lastly, the idea of reviewing courts writing reasons for administrators—putting words in their mouths—sits uneasily with the concept of deference to administrators.

Conclusion concerning the post-Dunsmuir cases

So the post-Dunsmuir jurisprudence is troubled and confused.

But in ascribing a cause to this, one must not cast the net too widely. Judges, even the most learned and senior ones around, cannot develop consistent and coherent jurisprudence unless they are guided by settled doctrine and well-accepted principles.

For that, Dunsmuir gets its share of blame. But, alas, in this area of law this failure has been with us for decades. Dunsmuir is just one manifestation of a wider problem, the failure to ground our operational rules in this area of law in settled doctrine and well-accepted principles.

What of Dunsmuir?

Dunsmuir was intended to bring order to chaos in this area of law. But, consisting of a series of decrees without doctrinal or principled basis, Dunsmuir has worsened the chaos and made the work of lower court judges like me more difficult.

For a moment, spare a thought for the judges. Put yourselves in our shoes. We took an oath to decide cases in accordance with the law. We deeply care about abiding by our oath and being correct on the law. We respect the hierarchy of courts and are keen to apply Supreme Court pronouncements on the law.

But what is the law? Without a doctrinal or principled basis to operate upon, what are we to do? In the case of Supreme Court authorities that conflict, which should we follow? Are some inconsistent Supreme Court authorities, as Professor Daly says, best regarded as “signal” and others as “noise” and so, despite our commitment to the hierarchy of courts and our deep wish to obey the Supreme Court, we should ignore some of its authorities? And in conducting reasonableness review—to paraphrase what Justices Moldaver and Wagner (as he then was) said in Kanthasamy at para. 112—should we do what the Supreme Court says or should we do what the Supreme Court does?

Almost every contributor to the Dunsmuir Decade series has criticized the current state of the jurisprudence, some quite severely. The number of critics and, in many cases, the seriousness of the criticism is really telling!

Even more telling is that, to a lesser or greater extent, all justices of the Supreme Court have departed from Dunsmuir though they don’t expressly say it—yet. One can see a significant departure in the Supreme Court engaging in variable intensities of review under the reasonableness standard, i.e., fussy review or not-so-fussy review, in many cases. Another is its relatively consistent tendency to conduct correctness review under the reasonableness standard. Most recently, very much at odds with Dunsmuir, the majority in Edmonton East casts doubt on whether contextual approaches can ever oust the presumption of reasonableness review for administrative interpretations of home statutes.

In fact, Dunsmuir causes such chaos that sometimes the Supreme Court itself avoids it completely in circumstances where it should be discussing it: see, e.g., Febles, B010, Tran. Its silence on Dunsmuir and the standard of review in these cases speaks loudly.

Not a single court in a Westminster jurisdiction has referred to Dunsmuir with praise or approval. These countries’ courts usually act doctrinally. Apparently Dunsmuir has nothing to teach them, nor does it have anything that is worthy of commendation or emulation.

This area of law is so important. It affects a large swath of executive government and those regulated by it. It mediates the relationship between executive government and the judiciary. Executive government serves the people and this area of law ensures that the people are treated legally and appropriately, with their rights protected. But inconsistency and incoherence gnaws at these objectives and brings this important area of law into disrepute.

Dunsmuir must be shown the door.

But what should replace it? That is the topic of my next article.

This content has been updated on March 7, 2018 at 13:50.