How the Lower Courts are “doing Dunsmuir” (Diana Ginn and William Lahey)

Diana Ginn is a Full Professor at he Schulich School of Law and William Lahey is President of the University of Kings College, and an Associate Professor, on leave, at the Schulich School of Law

The implications of Dunsmuir[1] for judicial deference towards administrative decision making were uncertain for at least two reasons.  The first was the elimination of review for patent unreasonableness.  The second was the more general call of the Supreme Court for less focus on the “esoteric” task of deciding whether or not deference was warranted and more on the question of importance to the parties involved: whether the administrative decision should be upheld. Dunsmuir called for this to happen through the increased use of precedent to determine whether an administrator’s decision on a legal issue[2] should be treated deferentially or not, and though the identification of four kinds of legal questions to which the correctness standard would automatically apply [3]– in either case, no standard of review analysis would be needed. Even where neither of these shortcuts was available, a consideration of one or more factors might be sufficient to determine whether to apply a reasonableness or correctness standard; a full four-part SORA was no longer required for every issue.

In order to determine whether, and if so how, lower courts are “doing Dunsmuir”, we[4] examined 400+ cases from across Canada (from Nova Scotia, Quebec, Ontario, Alberta and the federal courts, as well as from British Columbia, to compare how courts did substantive review under Dunsmuir and that province’s Administrative Tribunals Act[5]) between the date of Dunsmuir and 2015. Each decision was reviewed quantitatively against 100+ questions, and a significant number were reviewed further, to provide qualitative analysis. This research led to three articles, all published in the Canadian

Journal of Administrative Law and Practice.[6] It followed from the research we had completed in 2004[7] – based on reading approximately 300 cases from across Canada – where we examined how courts were understanding and implementing the framework established by Pushanathan[8]and Baker[9].  Then, we concluded that the courts had broadly absorbed the curial deference called for in CUPE v. NB Liquor Corp.[10]  On that basis we had, before Dunsmuir, advocated for replacement of two deferential standards of review with a single standard of reasonableness review both because the complexity of two difficult-to-distinguish deferential standards was unnecessary to ensure deference, and because more of judicial review should be concerned with applying rather than with choosing the standard of review.[11]

In our recent research, we discovered that, generally, courts do spend considerably more time on applying the standard of review than on selecting it – a change from what we observed in our earlier research. While courts used the categories of automatic correctness sparingly (for just over 10% of the issues reviewed fell), they accepted with alacrity the invitation to rely on precedent: the SoR was established by precedent for 64% of the issues before the federal courts, and for 67% of issues reviewed in Nova Scotia, Quebec, Ontario and Alberta. The factors most relied upon in determining the applicability of precedent were first, nature of the question and, second, identity of the decision-maker.

The courts in our study were also willing to choose standard of review based on a partial SoRA, although there was significant variation across the jurisdictions.  The federal court chose the SoR based on only one factor (generally nature of the question) for 68% of the issues it reviewed while courts in Nova Scotia, Quebec, Ontario and Alberta did so for 28% of the issues they reviewed. We found little correlation between whether a full or partial SoRA was performed and the standard selected.

This economy in the SoRA has not come at the expense of deference.  In fact, courts are more likely now than before Dunsmuir to apply deference – on average, 82% of the time in Nova Scotia, Quebec, Ontario and Alberta and 78% overall, compared to 70% in our earlier study. Further, courts are now more likely than they were before Dunsmuir to uphold the decision – on average for 71% of the issues reviewed in Nova Scotia, Quebec, Ontario, Alberta and the federal courts, as compared to in 66% of the cases reviewed in our earlier study. When immigration and refugee cases decided by the federal courts are removed from the calculations, issues were reviewed deferentially 82% of the time, and decisions were upheld 76% of the time. The numbers for British Columbia were very similar, suggesting a convergence between outcomes under Dunsmuir and B.C.’s Administrative Tribunals Act.  We also found that where courts did extensive analysis in reviewing for reasonableness, the decision was generally upheld.  It was not review for “correctness in disguise”.

Given that the premise of substantive review must be that a certain percentage of administrative decisions will be unreasonable or incorrect, this strikes us an impressive level of deference to administrative decision making.  Our sample of over 400 cases indicates that lower courts are “doing Dunsmuir” with remarkable fidelity, and that the deference which was already firmly established in the Pushpanathan-Baker era has been solidified and expanded. Courts are selecting the applicable standard of review in a more streamlined manner, and are working with only one loosely defined deferential standard or review, without being tempted to meddle unnecessarily with administrative action. Indeed, they have become more deferential.  We completed our third article in the series with this conclusion:

Dunsmuir gave the law of substantive review the revised and simplified framework it long needed more or less when the law was ready – and could be trusted – to put it to good, effective and proper use. All things considered, given the tortured history of our substantive review law since the 1960’s, Dunsmuir stands as an impressive accomplishment of law reform by the Supreme Court of Canada, and the jurisprudence of the lower courts stands as an impressive example of lower courts putting into practice those law reform principles.[12]

[1] 2008 SCC 9, [2008] 1 SCR 190

[2] By this point, the only real debate was around questions of law. Questions of fact and the exercise of discretion were accepted as attracting deference.

[3]These are: constitutional issues, “true questions of jurisdiction or vires”, “[q]uestions regarding the jurisdictional lines between two or more competing specialized tribunals”, and any question of general law “that is both of central importance to the legal system as a whole and outside the adjudicator’s specialized area of expertise”. (Dunsmuir, para 58-62)

[4] We are extremely grateful to our research assistants, who also became our co-authors (see note 6),  as well as to the Foundation for Legal Research and the Schulich School of law for financial assistance.

[5] S.B.C. 2004, c. 45

[6]  Ginn, Lahey, Soubolsky and Veinotte, “How has Dunsmuir Worked? A Legal-Empirical Analysis of Substantive Review of Administrative Decisions after Dunsmuir v. New Brunswick: Findings from the Federal Courts” (2017) 30 Can J Admin L & Prac 51;

Ginn, Lahey and Veinotte, “A Legal-Empirical Analysis of Substantive Review: Findings from the British Columbia Courts” (2017) 30 Can J Admin L & Prac 173; and

Lahey, Ginn, Constantine and Hooper, “How Has Dunsmuir Worked? A Legal-Empirical Analysis of Substantive review of Administrative Decisions after Dunsmuir v. New Brunswick: Findings from the Courts of Nova Scotia, Quebec, Ontario and Alberta” (2017) 30 Can J Admin L & Prac 317.

[7] Lahey and Ginn, “After the Revolution: Being Pragmatic and Functional in Canada’s Trial Courts and Courts of Appeal” (2002) 25 Dal LJ 259

[8] 1998] 1 SCR 982, [1998] SCJ No 46

[9] [1999] 2 SCR 817, [1999] SCJ No 39

[10] [1979] 2 SCR 227

[11] Supra note 7 at p. 328-29

[12] Supra note 6 at 346-7

This content has been updated on February 28, 2018 at 22:45.