Dunsmuir is Dead – Long Live Dunsmuir! An Argument for a Presumption of Correctness

By: Martin Olszynski

PDF Version: Dunsmuir is Dead – Long Live Dunsmuir! An Argument for a Presumption of Correctness

Case Commented On: Garneau Community League v Edmonton (City), 2017 ABCA 374 (CanLII)

Garneau is the latest judicial plea to the Supreme Court of Canada to do something about the standard of review – three judges, three judgments, all concurring in the result but each getting there somewhat differently. The case involves Alberta’s Municipal Governments Act, RSA 2000 c M-26, including statutory rights of appeal that are similar to those recently considered by the Supreme Court (and only slightly less recently considered by the Alberta Court of Appeal) in Edmonton (City) v Edmonton East (Capilano) Shopping Centres Ltd., 2016 SCC 47 (CanLII). In this post, I highlight Justice Watson’s and Slatter’s concerns about the standard of review framework as set out in Dunsmuir v New Brunswick, 2008 SCC 9 (CanLII) and its progeny. Before doing so, however, I first provide a primer on the Dunsmuir framework wherein I flag some of my own concerns. Drawing on these two parts, I then propose two concrete changes to the Dunsmuir framework that in my view would render it more coherent and stable, both doctrinally and practically. 

What has Dunsmuir Wrought?

Dunsmuir has stood principally for two things: first (and least controversially), the reduction of the number of standards of review from three to two (eliminating the standard of “patent unreasonableness”); and second, a purported simplification of what the Supreme Court now describes as the “Standard of Review analysis” (see Alice Woolley and Shaun Fluker, “What has Dunsmuir Taught?” (2010) 47 Alberta Law Review 1017, the title of which I have gratefully borrowed – with modification – for this part). Whereas before Dunsmuir courts and counsel were spending time and effort applying the “Pragmatic and Functional Approach”, discussing (i) privative clauses, (ii) the expertise of the tribunal, (iii) the purpose of the Act and the specific legislative provision(s) in play, and finally (iv) the nature of the question (see e.g. Pushpanathan v Canada (Minister of Citizenship and Immigration), [1998] 1 SCR 982, 1998 CanLII 778 (SCC)), Dunsmuir heralded the arrival of what was supposed to be a simplified inquiry (at para 62).

The first step is now to determine whether the standard of review has already been satisfactorily determined by past case law. When this question is answered in the negative, a court should then consider “whether the issue involves the interpretation by an administrative body of its own statute or statutes closely connected to its function” (Capilano at para 22). Where a decision-maker is interpreting their home statute, the reasonableness standard is presumed to apply. Likewise for questions of fact, mixed fact and law, and questions of discretion and policy (Dunsmuir at para 53). However, where a question of law falls into one of four correctness categories, the presumption is rebutted and correctness applies (Capilano at para 24): These are (i) constitutional questions regarding the division of powers; (ii) issues “both of central importance to the legal system as a whole and outside the adjudicator’s specialized area of expertise”; (iii) “true questions of jurisdiction or vires”; and (iv) issues “regarding the jurisdictional lines between two or more competing specialized tribunals”. On at least one occasion, the concurrent jurisdiction of a tribunal and the courts was sufficient to rebut the presumption of reasonableness (Rogers Communications Inc. v Society of Composers, Authors and Music Publishers of Canada, 2012 SCC 35 (CanLII)), but the Supreme Court has resisted making such statutory schemes their own correctness category. The presumption was also rebutted in Tervita Corp. v Canada (Commissioner of Competition), 2015 SCC 3 (CanLII) because the statutory right of appeal in that case was worded “as though originating from a court and not an administrative source” (at para 39).

Having applied Dunsmuir in my prior life as government counsel and having taught it now for a few years, I confess that it has seldom seemed like a simplification. Rather, as observed by Justice Abella in Wilson v Atomic Energy of Canada Ltd., 2016 SCC 29 (CanLII), it has felt mostly like a “labyrinth” (at para 19). It recently occurred to me that what my students needed, then, was a good road map, so I committed to laying out both the pre- and post-Dunsmuir frameworks in diagrammatic form.

Figure 1: The Pre-Dunsmuir “Functional and Pragmatic” Approach

Figure 2: The Dunsmuir Framework (per Edmonton East (Capilano) 2016)

Of course, even these two diagrams are an oversimplification. With respect to the “Pragmatic and Functional Approach”, Figure 1 does mask the nuance in the analysis, and the corresponding time devoted to it. With respect to Figure 2, Dunsmuir has resulted in a significant shift towards deference, such that the first step (whether there is a satisfactory precedent) has borne little fruit (indeed, this very problem was encountered in Garneau at para 9; see also Greenpeace Canada v Canada (Attorney General), 2014 FC 463 (CanLII) at paras 21-28). So it is usually necessary to proceed to the second step (the presumptions of reasonableness) but even that is not entirely accurate because applying the presumptions requires a court to first consider the nature of the question (as noted by Professors Woolley and Fluker, supra). I have taken to calling this step 1.5, wherein the courts (and counsel) reach into the “Pragmatic and Functional Approach”, extract what used to be the last step and make it the first substantive step.

Rebutting the presumption of reasonableness because the question of law is one to which correctness review applies is also not straightforward – even setting aside the thorny issue of “true questions of jurisdiction” (Alberta (Information and Privacy Commissioner) v Alberta Teachers’ Association, 2011 SCC 61 (CanLII)). I am referring to the category of issues “both of central importance to the legal system as a whole and outside the adjudicator’s specialized area of expertise” (emphasis added). Its consideration requires courts and counsel to reach once again into the “Pragmatic and Functional Approach”, this time to extract what used to be the second factor – determining the tribunal’s expertise.

Viewed this way, Figure 2 could be reworked as follows:

Figure 3: Dunsmuir De-constructed

Labyrinth indeed. What Figure 3 also makes clear is that the standard of review analysis has not been so much simplified as reorganized, with privative clauses and the purpose of the legislative scheme relegated to tertiary (quaternary?) status. For a doctrine that is supposed to reconcile the rule of law with legislative supremacy, this result is puzzling.

And while I may be guilty of exaggerating the complexity a little bit (there are certainly many cases where the analysis ends at the 2^nd step; for a recent Alberta example, see e.g. Bow River Irrigation District v Wilkinson, 2017 ABQB 616 (CanLII)), I am not exaggerating my students’ frustration with what seems to be a hopelessly inchoate area of law. Such sentiments were put on full display when I asked my students to prepare their own diagrammatic representations of Dunsmuir. The following one (which I received permission to use) struck me as particularly representative of this area of law, especially where it asks “Do you feel like doing more analysis?”

Figure 4: A Student’s Diagrammatic Representation of Dunsmuir

Garneau: “The day has [still] not yet come…”

As noted above, Garneau involved provisions of Alberta’s Municipal Governments Act (MGA). The Edmonton Subdivision and Development Appeal Board (SDAB) had overturned a Development Officer’s refusal to grant a development permit for the construction of an apartment building in Edmonton. An appeal of the SDAB’s decision was filed by the Garneau Community League, permission for which could be granted if it “involves a question of law of sufficient importance to merit a further appeal and has a reasonable chance of success” (MGA s 688(3)). Prior to the Supreme Court’s decision in Capilano, which considered similar provisions (MGA s 470(1), (5)), the jurisprudence had established correctness as the appropriate standard of review for decisions of the SDAB (at para 9). The question before the Court of Appeal was whether the Supreme Court’s selection of reasonableness review in Capilano required a similar result in Garneau.

Madam Justice Strekaf decided that determining the standard of review was unnecessary because the Board was correct with respect to one of the issues, and both unreasonable and incorrect with respect to the other (at para 12). Mr. Justice Watson concurred in the result, but felt compelled to provide his own perspective on the standard of review. In his view, Capilano was not determinative of the standard of review issue. Justice Watson reviewed the case law in which Dunsmuir’s presumption of reasonableness was rebutted in situations other than the four established correctness categories (at para 55-57) and concluded that correctness ought to be applied to the SDAB’s decisions. In his view, such a result was more consistent with the legislature’s intent (i.e. given the Court of Appeal’s long-established practice of applying the correctness standard, the legislature could be taken to have acquiesced to that standard) (at para 65). Reiterating that “judicial review has been linked to the constitutionally recognized rule of law role of Courts to evaluate and determine the legality of an action of any state agency or state agent” (at para 69), Justice Watson concluded that if there were to be movement in standard of review doctrine, “that movement should be towards a method of identification of what are the extricable questions of law or jurisdiction for which Parliament or a Legislature or the common law would expect or demand a single answer” (at para 73).

Finally, Mr. Justice Slatter also provided concurring reasons, concluding that correctness applied primarily because the relevant regime involved multiple decision-makers: “…it is untenable that planning legislation means one thing in one municipality and something else in another. The standard of review is accordingly correctness” (at para 77). Justice Slatter goes on to suggest that the confusion may be rooted in the core of the Dunsmuir framework. At the same time, however, Justice Slatter appears to be of the view that the answer lies in Dunsmuir itself:

[91] What is judicial review supposed to be? Some comments from Dunsmuir may show the way forward:

[27] As a matter of constitutional law, judicial review is intimately connected with the preservation of the rule of law. It is essentially that constitutional foundation which explains the purpose of judicial review and guides its function and operation. Judicial review seeks to address an underlying tension between the rule of law and the foundational democratic principle, which finds an expression in the initiatives of Parliament and legislatures to create various administrative bodies and endow them with broad powers. Courts, while exercising their constitutional functions of judicial review, must be sensitive not only to the need to uphold the rule of law, but also to the necessity of avoiding undue interference with the discharge of administrative functions in respect of the matters delegated to administrative bodies by Parliament and legislatures.

[28]  […] Judicial review is the means by which the courts supervise those who exercise statutory powers, to ensure that they do not overstep their legal authority. The function of judicial review is therefore to ensure the legality, the reasonableness and the fairness of the administrative process and its outcomes. …

[30]   […] In essence, the rule of law is maintained because the courts have the last word on jurisdiction, and legislative supremacy is assured because determining the applicable standard of review is accomplished by establishing legislative intent.

If this is not the purpose of judicial review, what is it? If “legislative intent” is the key, then why does the express provision of a right of appeal to an appellate court make no difference to the analysis? […]

Both Justice Watson and Justice Slatter appear to reaffirm the constitutionalization of judicial review, which has its roots in the Supreme Court’s decision in Crevier v Attorney General of Quebec, 1981 CanLII 30 (SCC). Justice Slatter also recognizes the importance of legislative supremacy but is bewildered by the privileging of implied intent over explicit intent. As further discussed below, they are not alone in these sentiments. Before leaving Justice Slatter’s reasons in Garneau, however, it is useful to cite two final paragraphs. The first reflects some of the concerns raised in the first part of this post, while the second speaks to the problems with “expertise” as a basis for deference:

[93]  Rhetoric about deference has overtaken the analysis, and too frequently an attempt is made to fix the standard of review without remembering that “standard of review” is merely a means to an end, not an end itself. Rigidity has appeared in the Dunsmuir analysis. The rebuttable presumptions in Dunsmuir, such as the one that interpretations of the tribunal’s home statute are reviewed for reasonableness, are turning into conclusive presumptions. Capilano Shopping Centres at para. 32…accepted that the presumption of reasonableness can be rebutted if the legislative context shows correctness was intended, but refused to apply that concept in a most obvious circumstance. Dunsmuir’s four categories calling for a correctness review are shrinking, and any attempt to identify new categories is resisted without reference back to the underlying principles. Even attempts to refine categories such as “questions of law of importance to the legal system” and “interpretation of the home statute” are strongly resisted.

[94]  The rhetoric about deference is perhaps matched by the rhetoric about “expertise”. Administrative decision makers undoubtedly develop an expertise, but the interpretation of statutes is not at the core of tribunal expertise. The superior courts have generally greater, but at least equal expertise in interpreting statutes. When the Legislature provides a direct appeal to the superior courts, does that not signal a recognition of the superior court’s expertise on questions of law, and also invite the exercise of that expertise by the superior courts? […] 

Re-building Dunsmuir

As noted above, there is once again in Canada growing dissatisfaction with the state of administrative law doctrine. In his article, “The Canadian Law of Judicial Review: A Plea for Doctrinal Coherence and Consistency”, Justice Stratas of the Federal Court of Appeal expressed views similar to those of Justices Watson and Slatter: “…the Supreme Court in Dunsmuir planted the right seeds and initially did much to help them germinate. Dunsmuir is doctrinally sound. But…the Supreme Court has allowed weeds to grow in the garden, choking and obscuring what ought to be thriving and clear” (at 13). He has also expressed the same bewilderment as Justice Slatter about the subservience of explicit legislative intent to implied intent (at 4 and 5; citations omitted):

Legislation sometimes signals that the standard of review should be correctness—no deference at all to the administrative decision-maker. In some cases, the Supreme Court reads these signals and properly carries out the legislator’s intent, reviewing the decision for correctness.

But sometimes not… Even where the legislator has granted a full right of appeal, there is a presumption that administrative interpretations of legislation are subject to deferential reasonableness review.

In my view, resolving these tensions requires two inter-related changes to the Dunsmuir framework that, at first blush, may seem drastic but that upon closer examination should preserve much of the post-Dunsmuir case law. The first change would be to reverse the presumption of reasonableness on questions of law to a presumption of correctness, which can then be rebutted for the large majority of such questions through the presence of a privative clause (this approach would be similar to that proposed by Justice Deschamps in Dunsmuir). The second related change would be to abandon the overly broad and fundamentally contradictory concept of “expertise” as a basis for deference and to replace it with the potential for democratic accountability, which ultimately is the basis for legislative supremacy.

As a starting point, it is useful to acknowledge that we are concerned only with the standard of review applicable to questions of law. The application of deference and the standard of reasonableness to questions of fact and questions of mixed fact and law is relatively well settled (H.L. v Canada (Attorney General), 2005 SCC 25 (CanLII), as cited by Deschamps J in Dunsmuir at para 161).

In my view, Dunsmuir’s presumption of deference with respect to questions of law is fundamentally backward. If judicial review is linked to the courts’ constitutional role to determine the legality of state action, then any presumptions should reflect the courts’ basic core competency, i.e. the interpretation of law. As recently described by the Supreme Court in Ontario v Criminal Lawyers’ Association of Ontario, 2013 SCC 43 (CanLII):

[28]  Over several centuries of transformation and conflict, the English system evolved from one in which power was centralized in the Crown to one in which the powers of the state were exercised by way of distinct organs with separate functions. The development of separate executive, legislative and judicial functions has allowed for the evolution of certain core competencies in the various institutions vested with these functions.  The legislative branch makes policy choices, adopts laws and holds the purse strings of government, as only it can authorize the spending of public funds. The executive implements and administers those policy choices and laws with the assistance of a professional public service. The judiciary maintains the rule of law, by interpreting and applying these laws through the independent and impartial adjudication of references and disputes, and protects the fundamental liberties and freedoms guaranteed under the Charter.

(See also Justice Mainville’s decision in Canada (Fisheries and Oceans) v David Suzuki Foundation, 2012 FCA 40 (CanLII) [Killer Whale])

Simply put, the starting point should be that the courts, by virtue of their training, independence, and impartiality, have the upper hand in the interpretation of the law. Recognizing the realities of the modern administrative state, however, this presumption can and should be rebuttable for certain questions of law by virtue of explicit legislative provisions (i.e. privative clauses and restrictive rights of appeal). Importantly, just as the Supreme Court in Crevier held that legislatures could not oust judicial review entirely, so too certain questions of law will always be subject to correctness review – these would be the current Dunsmuir correctness categories (more or less, as further discussed below). For all other questions of law, however, the presence of a privative clause would trigger deferential review.

In addition to better reflecting the conventional separation of powers, such an approach is also preferable from the perspective of legislative supremacy. Why do legislatures matter? They matter – and must be respected – because they are democratically elected and accountable. Democratic accountability is hindered, however, when the unelected courts rely on implied legislative intent in a blanket manner regardless of context. An analogy can be made to the defenses of statutory authorization and statutory immunity in the context of nuisance law. In the case of statutory authorization, courts engage in a complex and policy-oriented inquiry to determine whether a nuisance, e.g. one associated with the construction of British Columbia’s Canada Line, was implicitly authorized by virtue of granting one or more relevant statutory permits (see e.g. Susan Heyes Inc. (Hazel & Co.) v South Coast B.C. Transportation Authority, 2011 BCCA 77 (CanLII)). In the case of statutory immunity, no such inquiry is required because the legislature has been explicit, as has been the case with respect to various polluting industries throughout Canada (Jamie Benidickson, Environmental Law, 4^th ed (Irwin Law: Toronto, 2013) at 105-106). At the risk of stating the obvious, it is much easier to hold legislatures accountable for their explicit choices than for their judicially inferred ones (as Professor Donal Nolan wisely observes, statutory authority is actually “a judicial, as opposed to a statutory defense, since the authority for the proposition…comes from case law rather than legislation”: “Nuisance, Planning and Regulation: The Limits of Statutory Authority” in Andrew Dyson, James Goudkamp and Frederick Wilmot-Smith (eds), Defences in Tort (Hart Publishing 2015) at 186). Similarly, it is much easier to hold legislatures accountable for modifications to the conventional separation of powers (e.g. to expand the executive’s role in legal interpretation) when they have to be explicit about it.

Although privative clauses and rights of appeal may seem mundane or overly technical to some observers, they do matter and can be politically contentious depending on the context. It is probably true, on the one hand, that few labour or employment lawyers would argue against privative clauses in that context, and indeed privative clauses enjoy a long history there. They do not, on the other hand, have a long history in the environmental law context, and their inclusion here can indeed be controversial, bearing in mind that it was governments’ miserable record of taking environmental concerns into account that spurred the development of environmental laws in the first place. Simply put, if Parliament or the legislatures want courts to defer to government agencies on interpretations of law, an arrangement that can be easily understood by all Canadians as a departure from the conventional separation of powers, then it makes sense that they should have to “take it on the chin” – to do so explicitly and be prepared to be held accountable for it (whether in Parliament, in Committee Hearings, or at the ballot box).

For these reasons, it is not sufficient to argue, as I suspect some Dunsmuir defenders might, that such a choice was made along with the choice “to delegate decision making to a tribunal, rather than the courts” (Capilano at para 22). It is also incorrect, because it lumps into a single group those situations where indeed Parliament or the legislatures have deliberately removed certain matters from direct judicial involvement (e.g. labour and employment disputes, human rights complaints, and even some environmental matters (e.g. those subject to appeal to Alberta’s Environmental Appeals Board)) and those in which the courts never played any direct role. It makes no sense to say, for example, that the courts should defer to the Minister of Fisheries and Oceans in his or her interpretation of the Fisheries Act or the Species at Risk Act because the legislature “chose the Minister rather than the courts” to make decisions under those statutes – the courts never played such a role. Rather, the courts’ well-established role in this context has been to determine whether executive decision-making has been “exercised within the legal framework provided by the legislation” (Alberta Wilderness Association v Canada (Attorney General), 2013 FCA 190 (CanLII) at para 48).

An insistence on explicit legislative provisions and the associated potential for democratic accountability would have one additional – and critical – salutary effect: the abandonment of implied expertise as the other basis for deference. According to the Supreme Court in Capilano, “The presumption of reasonableness is grounded in the legislature’s choice…and the expertise of the tribunal… Expertise arises from the specialization of functions of administrative tribunals like the Board which have a habitual familiarity with the legislative scheme they administer…” (Capilano at 33, citing Dunsmuir). In my view and with respect, this concept is too contradictory and naive to stand as a basis for deference on questions of law.

As understood in the Canadian administrative law sense, the contradictory nature of expertise is well captured by the following photograph (entitled “Pushpumkinathan”) that found its way into my twitter feed after I first shared some of the diagrams contained in this post (with thanks to Elin Sigurdson):

The pumpkin on the left is “correct”, the one in the middle is “reasonable”, while the one on the right is “patently unreasonable”. Another analogy used by University of Ottawa Professor Craig Forcese (with credit to Professor Jamie Benidickson) is that of a dartboard: “Correctness requires the tribunal to hit the bull’s eye; reasonableness require[s] the tribunal to hit to dartboard.” And yet courts are told to defer to these less-than-perfect pumpkin carvers and dart players on the basis that they are expert pumpkin carvers and dart players.

In my view, this contradiction can be explained by acknowledging that it is not solely “expertise” in the traditional sense (e.g. “having, involving, or displaying special skill or knowledge derived from training or experience” – according to Merriam Webster) that drives administrative decision-making. The Supreme Court appears to have acknowledged this reality when it has at times equated expertise with “field sensitivity” (Dunsmuir at para 49), but neither term captures the reality of the situation: that administrative decision-making is animated by several, often competing, forces including not only subject-matter expertise but also resource constraints, individual values and preferences, and political pressure(s), and only some of which will be authorized by the relevant statute. Perhaps more fundamentally, this whole conversation assumes that administrative decision-makers’ primary concern is legal interpretation when in reality it is often the substantive outcome that matters just as much – and sometimes more. In those instances, reasonableness review on questions of law provides some room to maneuver; exactly how much being the stuff of litigation risk matrices and management.

I concede that courts are not immune from all of these other forces; individual preferences and values can and do influence judicial analysis, as can a concern for substantive outcomes. But it is untenable to conflate the pushes and pulls of administrative decision-making with that of the courts, where independence and impartiality are the guideposts and legal interpretation is the primary exercise. And to reiterate, I am not suggesting that deference should never be granted on questions of law, but only that “implied expertise” or “field sensitivity” is an inadequate basis for doing so. The fact that a legislature has been explicit about demanding deference and can be held accountable for it (the approach argued for here) is at least a better basis.

It also clarifies the nature of the reasonableness analysis, which in turn should further enable that accountability. Reasonableness is best understood as the courts saying, “close enough” – with the knowledge that a decision-maker may not be “correct” in their interpretation – rather than “we do not have the expertise to fully grasp this question or its answer.” To paraphrase the Supreme Court, if expert tribunals cannot explain the reasons for their conclusions, they are not very expert (Canada (Director of Investigation and Research, Competition Act) v Southam Inc., 1997 CanLII 385 (SCC) at para 62). Indeed, it is my sense that where an interpretation of law is genuinely rooted in a decision-maker’s expertise (e.g. in competition law, utilities regulation etc.), they will generally be in least need of deference because their interpretations will also probably be quite compelling. It is where their decisions appear to run counter to such expertise (as was the case in Killer Whale) that administrative decision-makers are on shakier ground, and where the true stakes of judicial deference become plain.

In sum, and because it was diagrams that got me here in the first place, a revised approach to determining the standard of review could look like this:

Figure 5: A Revised Dunsmuir Framework

Perhaps it is still not as tidy as I (or my students) would have hoped, but I do think that this proposed approach better reflects (i) the constitutional role of the courts and judicial review in particular; (ii) the principle of legislative supremacy; and (iii) the nature of deference in the context of reasonableness review.

Before concluding this post, I want to flag some unaddressed questions that will require further thought. The first is the appropriate approach to Charter issues. The current approach was laid out in Doré v Barreau du Québec, 2012 SCC 12 (CanLII) and Loyola High School v Québec, 2015 SCC 12 (CanLII). The Ontario Court of Appeal recently cast some doubt on this approach in E.T. v Hamilton-Wentworth District School Board, 2017 ONCA 893 (CanLII) (at paras 108 – 125), concluding that it “is one thing to defer to an educator on educational matters, but something else to defer to an educator on constitutional matters.” The second question has to do with the correctness category concerning “questions of central importance to the legal system”. In light of the discussion above with respect to expertise, it would probably not surprise anyone if I suggested that the second part of this inquiry (whether the question falls outside the tribunal’s expertise) should be abandoned. However, this too requires some further thinking on my part.

I am grateful to Professors Nigel Bankes, Howard Kislowicz and Jennifer Koshan for their thoughtful comments on an earlier version of this post.

This post may be cited as: Martin Olszynski “Dunsmuir is Dead – Long Live Dunsmuir! An Argument for a Presumption of Correctness” (13 December, 2017), online: ABlawg, http://ablawg.ca/wp-content/uploads/2017/12/Blog_MO_Garneau.pdf

To subscribe to ABlawg by email or RSS feed, please go to http://ablawg.ca

Follow us on Twitter @ABlawg