Dunsmuir – Plus ça change
On Friday, March 7, 2008, the Supreme Court of Canada released Dunsmuir v New Brunswick,  1 SCR 190 [“Dunsmuir”], a stark reversal of the last decade of administrative law jurisprudence on the issue of the standard of review. This decision, while undoubtedly a landmark judgment, may change little about how courts review administrative action. I explore the reasons for this assessment below.
I should add that the Dunsmuir decision also reverses a longstanding principle of the application of procedural fairness to public sector labour employment settings. The focus of this brief comment, however, is the standard of review by which courts review administrative action.
Lebel and Bastarache JJ., begin their majority reasons in Dunsmuir by throwing down the following gauntlet:
This appeal calls on the Court to consider, once again, the troubling question of the approach to be taken in judicial review of decisions of administrative tribunals. The recent history of judicial review in Canada has been marked by ebbs and flows of deference, confounding tests and new words for old problems, but no solutions that provide real guidance for litigants, counsel, administrative decision-makers or judicial review judges. The time has arrived for a reassessment of the question.
With this opening, the actual facts of Dunsmuir appear almost an afterthought, more an excuse for the development of a doctrine, than the driver of the decision. This is something of an irony because in so many standard of review cases, it is the facts that seem to drive the doctrine rather than the other way around.
That said, here in a nutshell is what brought the case to the Supreme Court. Mr. Dunsmuir was a civil servant employed in the Department of Justice in New Brunswick. The Government terminated his employment, due to a number of problems, and offered him severance in lieu of notice. Mr. Dunsmuir took the position that he was entitled not just to a contractual remedy but also to fairness before being terminated (e.g. an opportunity to know the concerns of the employer and a chance to address them). A labour arbitrator was appointed to address Mr. Dunsmuir’s challenge and concluded that fairness was indeed required, and had not been provided. The reviewing judge reversed the arbitrator’s finding, finding at various points that the applicable standard of review was correctness, reasonableness simpliciter and patent unreasonableness. The notion that a single decision by a single decision-maker would require a court to parse through three different standards of review vividly illustrates the complexity that the Supreme Court then sets out to remedy. The Court of Appeal for New Brunswick dismissed the appeal from the reviewing judge, and the matter thus ended up at the Supreme Court.
Before proceeding to consider which path the Court opted for, it is important to be clear about the stakes in the standard of review debate.
The standard of review doctrine in administrative law asks when a court may overturn an administrative decision. On the one hand, this doctrine engages the rule of law. If parties cannot challenge administrative decisions before independent judges, then there is no guarantee that fundamental rights will be safeguarded. On the other hand, administrative decision-makers, especially adjudicative and regulatory tribunals, are created for the very purpose of providing an alternative to courts, and to respect the integrity of these statutory schemes, judges must show deference to the administrative decision-makers.
The standard of review thus represents the search for a constructive relationship between courts and administrative decision-makers which reflects respect for the rule of law, Parliamentary supremacy and the complex decision-making environments of the modern state.
In 1979, the Supreme Court ushered in the modern age of the standard of review in CUPE v New Brunswick Liquor,  2 SCR 227, in which the Court affirmed a new paradigm of review emphasizing deference. Dickson J. (as he then was) characterized this deferential standard as one of “patent unreasonableness,” under which the Court would only intervene in an administrative decision where that decision could not be rationally justified in light of the statutory authority of the decision-maker. Where this deferential standard did not apply, the Court applied the standard of “correctness”. In the 1988 case of Union des Employés de Service, Local 298 v Bibeault,  2 SCR 1048, the Court introduced the “pragmatic and functional” approach that would come to characterize the analysis by which the proper standard of review is determined (and included both a close reading of the relevant statutory provisions, but also a look at the broader purposes of the administrative scheme, and the specific expertise of the decision-maker.) Finally, in the 1996 case of Canada (Director of Investigation and Research) v Southam Inc,  1 SCR 748 , Iacobucci J., writing for the Court, introduced an intermediary standard of review, known both as “reasonableness simpliciter” and simply “reasonableness.”
The consensus on the three standards of review and the pragmatic and functional approach used to determine which was applicable was no sooner confirmed (in a unanimous judgment in 2003 in the case of Ryan v Law Society of New Brunswick,  1 SCR 247) than it began to unravel. In Toronto (City) v CUPE,  3 SCR 77, also a 2003 Supreme Court decision, Lebel J. (writing in a concurring judgment for himself and Deschamps J.) acknowledged the “growing criticism” and “serious questions” which had emerged over the standard of review jurisprudence of the Court. In particular, Lebel J. questioned whether a court could meaningfully distinguish between the “reasonableness simpliciter” and “patent unreasonableness” standards, and also questioned why this was necessary.
In Dunsmuir, Lebel and Bastarache JJ. refer to the whole enterprise of attempting to shore up the distinctiveness of the three previous standards, and the pragmatic and functional methodology, as having “proven difficult to implement.” Binnie J., writing concurring reasons in Dunsmuir, characterized the previous approach more bluntly as “distracting” and “unproductive.”
So, what has the Court developed in its stead?
First, the majority in Dunsmuir concluded that the distinction between “reasonableness” and “patent unreasonableness” was untenable and so, henceforth, there will only be two standards of review: reasonableness and correctness. While the standard of correctness is straightforward (a Court will intervene any time it concludes that an administrative decision-maker erred), the standard of reasonableness remains opaque (especially under the new scheme, under which the old standards of patent unreasonableness and reasonableness simpliciter have been collapsed into a single standard). In attempting to elaborate on this new, expansive standard of reasonableness, Lebel and Bastarache JJ. view the reasonableness analysis flowing from the “justification, transparency and intelligibility” of an administrative decision.
As for the approach to determining which standard of review is applicable, the majority has retained the substance of the pragmatic and functional approach (although it has now been repackaged as simply the “standard of review analysis”), but has noted that an exhaustive review is not always needed to determine the proper standard. Consequently, the Court has established both more flexible standards and a more flexible methodology for determining which standard applies to particular settings.
As the dust settles, what has changed? The majority is at pains to emphasize that the new scheme is not intended to pave the way for a more intrusive and less deferential approach to the review of administrative action. Lebel and Bastarache JJ. also confirm that deference continues to convey the idea of respect for the legislative choices of government.
Binnie J., in his concurring reasons, endorses the reappraisal of the judicial review system but stresses that it must extend beyond adjudicative tribunals, which are the focus of the majority, to all administrative decisions, including the decisions of those he describes memorably as “the lesser officials who reside in the bowels and recesses of government departments adjudicating pension benefits or the granting or withholding of licenses or municipal boards poring over budgets or allocating costs of local improvements.” After noting the diversity of the kinds of administrative decisions caught by the standard of review analysis, Binnie J. concludes that “‘reasonableness’ is a big tent…” As he notes, one significant change wrought by this decision is that courts will have to accommodate a range of deferential postures within the standard of reasonableness. This may simplify the analysis but render the job of the reviewing judge no less complex and potentially more opaque.
A further concurring set of reasons authored by Deschamps J. (writing for herself and Charron and Rothstein JJ.) seeks to simplify the standard of review further, by suggesting the primary criterion in the analysis should be the nature of the problem – that is, whether the decision under review was on a question of fact, mixed fact and law or law, with deference attaching to decisions on questions of fact, less deference attaching to decisions on questions of mixed fact and law and no deference attaching to decisions on questions of law.
While it is clear that the standard of review as a legal doctrine has been changed by Dunsmuir, it is not clear that any case previously decided on standard of review grounds would have resulted in a different outcome if the Dunsmuir framework were applied. This is in contrast to the aspect of Dunsmuir dealing with the applicability of procedural fairness to the decision to terminate public employees, where the Court reversed its earlier position in Knight v. Indian Head School Division, No. 19 (1990). Whereas a wide range of public sector labour cases which turned on fairness considerations will now potentially be decided differently, it is hard to point to a standard of review decision where the outcome might now be in doubt. Indeed, the majority and concurring judges in Dunsmuir apply different standards to the decision of the arbitrator but all conclude that the arbitrator’s decision cannot stand. In this sense, it would appear that not much has changed at all. Dunsmuir, I would suggest, is not about changing results, but rather about getting to the appropriate result more efficiently, more transparently and more coherently.
Dunsmuir is a step in the right direction. From the overly formalistic pragmatic and functional approach, we have now entered an era of what I would characterize as the contextual and transparent approach. The sometimes necessary but never tenable distinction between patent unreasonableness and reasonableness simpliciter has been happily abandoned. The dilemmas of complexity and inconsistency which plagued the standard of review analysis, however, likely have not been resolved. Courts will now puzzle over different degrees of deference within each standard and in what circumstances more or less exhaustive applications of the standard of review analysis might be appropriate.
Where does the standard of review analysis head from here? In “Contextual Snakes and Ladders: Iacobucci’s Legacy and the Standard of Review in Administrative Law” (2007) University of Toronto Law Journal 581, Colleen Flood and I argued that the Supreme Court should borrow a page from its successful procedural fairness jurisprudence in order to complete the “contextual turn” in the standard of review. In its approach to determining the degree of fairness applicable to a wide variety of administrative settings, the Court developed in Baker v. Canada (Minister of Citizenship and Immigration) (1999) a set of criteria (nature of the decision, nature of the statutory scheme, effect of the decision on the party, legitimate expectations of the affected party and the choice of procedure by the decision-maker) in order to place the obligation of fairness along a spectrum. For ease of reference, the Court has referred to a minimum, medium and high degree of fairness, but these clusters along the spectrum are not intended to represent fixed or rigid categories. The fairness spectrum allows reviewing courts to calibrate the degree of fairness applicable to the actual circumstances and facts of particular cases.
After reading Dunsmuir, I am more convinced than ever that a spectrum of deference approach will lead to greater transparency and a fuller embrace of the contextual analysis of deference first urged by Dickson J. (as he then was) in CUPE v. New Brunswick Liquor almost 30 years ago. The Court’s openness to revisit, revamp and reinvigorate the standard of review in Dunsmuir should be welcomed. The standard of review, however, remains a work in progress.