Alice Woolley, University of Calgary
I love criticizing Dunsmuir v New Brunswick. Who doesn’t? The majority judgment purports to simplify the identification of the standard of review, but sets out a test with the potential to bog a court down (precedent + the “standard of review analysis”). It suggests reasonableness as a deferential standard, yet invites courts to look not just at an administrative decision-maker’s reasons, but also at the outcome it reaches – an apparent invitation to a court to assess the substance of a case, instead of focusing on the administrative decision-maker’s analysis. Indeed, in Dunsmuir itself the majority neither simplified its own identification of the standard of review nor genuinely deferred to what the labour arbitrator in that case had decided (as David Mullan pointed out at the time). And of course perhaps the most telling criticism arises from the case’s failure to accomplish what one would have thought was its core mandate: allowing the Supreme Court to move on from its preoccupation with, and disagreement about, what is a ultimately a secondary question in any administrative law case. Standard of review is what the Supreme Court can’t quit. As recently as 2016 the Court split 5-4 on the appropriate standard of review in Edmonton (East) Capilano Shopping Centres Ltd.. And even in a straightforward case, the Court can still spill a remarkable amount of ink identifying the standard of review (see, e.g., Green v Law Society of Manitoba (2017)).
It thus saddens me to concede that this blog does not criticize Dunsmuir. It does not even offer praise only as cover for a nasty zinger or devastating critique. It instead reinforces the empirical studies by Gerald Heckman, Robert Danay and others (summarized by Paul Daly here) to suggest that, on the whole, the effect of Dunsmuir on judicial review of administrative decisions has been more positive than negative.
My contribution to the empirical conversation was to review Court of Appeal and Supreme Court decisions reported on CanLII involving judicial review of law society decisions (mostly through statutory rights of appeal). I reviewed 76 cases with “law /1 society” in the title and “standard /2 review” in the text, identifying the 59 decisions involving judicial review (the remaining cases were disputes to which the law society was a party). Of those, 40 cases were decided after Dunsmuir, and 19 before. I chose law society decisions because I can read those decisions and understand the issues quickly, which allows assessment of, for example, whether the law society acted badly such that the court’s interference was understandable, or whether the court was really just substituting its judgment for the law society’s. I analyzed the cases on the following grounds:
- Did the decision uphold or reverse the law society?
- Was the decision able to identify standard of review in five paragraphs or less?
- Was the standard identified reasonableness or correctness?
- Was the Court’s judgment analytically weird (for example, the Adams v LSA (2000) where the Alberta Court of Appeal applied the “error of principle or… unreasonable or demonstrably wrong” standard of review (and concluded that the decision was both “correct” and not “manifestly unreasonable”))
- Did the court in fact defer taking into account both the analytical methodology (i.e., focusing on the reasons of the law society rather than on the court’s own analysis of the issue in the case) and the grounds supporting the court’s interference with the law society’s decision?
My analysis suggests little difference between how willing courts are to uphold law society decisions before and after Dunsmuir. In both time periods courts upheld the law society decision more than 75% of the time (75% post-Dunsmuir; 79% pre-Dunsmuir). Nor are judges notably more willing to use reasonableness review after Dunsmuir than they were before, at least on some of the issues raised by a case (84% pre-Dunsmuir; 90% post-Dunsmuir, including two dissenting judgments).
More significant differences arise with respect to how often the court is willing to use correctness, the length of its analysis in identifying the standard of review, and the likelihood of an odd judgment. In the pre-Dunsmuir cases the courts used correctness for at least part of the decision regularly – in 7 of the 19 cases (37%) correctness was employed in whole or in part. Conversely, in only 2 of the 40 post-Dunsmuir decisions was correctness used (5%), with one additional dissenting judgment using correctness (bringing the total to 7%).
Similarly, prior to Dunsmuir, courts regularly devoted a considerable amount of their decisions to identifying the standard of review. In the pre-Dunsmuir cases 32% of judgments spent more than 5 paragraphs identifying the standard of review, whereas subsequent to Dunsmuir only 12% did. In many of the post-Dunsmuir cases the court identifies the standard of review in a single paragraph.
Further, in almost all of the post-Dunsmuir cases the standard of review used by the court recognizably conformed to the Supreme Court’s jurisprudence. In one post-Dunsmuir case, DeMaria v LSS (2015) the Saskatchewan Court of Appeal referenced the idea of a court needing a “palpable and overriding error” for reviewing a tribunal’s findings of fact, relying on Justice Deschamps’ concurring judgment in Dunsmuir, but the Court connected that concept to the idea of reasonableness, which seems defensible. In another British Columbia decision, Mohan v LSBC (2013), the Court did not identify the standard of review, but appeared to review for correctness a Law Society Review Panel’s assessment of a Hearing Panel decision, in circumstances where the appropriateness of a correctness standard is not obvious (the issue before the Court was the Review Panel’s assessment of the Hearing Panel’s findings of fact). The Court’s failure in that case to engage at all with its own standard of review is unusual. Mohan is the one post-Dunsmuir law society decision where, to my mind, the Court dropped the ball entirely on the standard of review analysis.
In the pre-Dunsmuir cases courts also mostly seemed to grasp the basic concepts that govern standard of review (correctness, reasonableness and (at the time) patent unreasonableness) but in 3 of the 19 judgments the court’s analysis seemed odd or off-base, which is a much higher percentage (16%) than the 2.5% (1/40) for the post-Dunsmuir decisions. In its 2000 judgment in Adams, referenced earlier, the Alberta Court of Appeal treated an error of principle, unreasonableness and being demonstrably wrong as a singular standard of review, and later upheld the law society’s decision on the basis that it was both correct and manifestly reasonable. The Court did not obviously appreciate the basic ideas of standard of review and judicial deference, or turn its mind to the level of deference it needed to exercise. In a 1999 decision, Phillon v LSA, the Alberta Court of Appeal did not identify the standard of review explicitly but upheld the law society in part “given the standard of review” – implying, I think, a reasonableness standard. Yet the Court also overturned part of the law society decision on the basis that the law society had sanctioned a lawyer using guidelines not published until after the hearing. For that aspect of the decision the Court did not identify the standard of review it was using. That standard on that part of the decision appears to have been correctness but the court never says so. In 1993 the British Columbia Court of Appeal in McOuat v LSBC suggested that courts may only interfere with administrative decisions that fall within the decision-maker’s jurisdiction where the decision-maker has “abused its discretion”, which seems to adopt the pre-CUPE idea of radically different levels of deference based on the administrative decision-maker’s jurisdiction, rather than a post-CUPE concept of judicial review.
To my mind this comparison speaks in Dunsmuir’s favour. Some of these effects almost certainly arise from more than that decision – it is likely that less odd judgments arise post-Dunsmuir because of the Supreme Court’s repeated reiteration of the concepts of standard of review, not Dunsmuir on its own. A judge has to be pretty out of it nowadays not to know that correctness and reasonableness are the core concepts of standard of review, and to miss the need to identify the standard of review when considering an administrative decision. Nonetheless, the reduction in odd judgments is a good thing, and Dunsmuir deserves at least some of the credit for it. Further, the increased simplicity of identifying the standard of review is great. In almost all law society cases reasonableness should be the standard – so why belabor the identification of it? That simplification was something Dunsmuir explicitly sought to accomplish, and these cases suggest that it has done so. Dunsmuir also does not seem to have made courts any more willing to interfere in law society decision-making – courts, at least when it comes to law societies, are respectful of and deferential to administrative authority, and Dunsmuir has made them no less so.
Indeed, perhaps the thing that strikes me most about reading these judgments is how sound the courts’ instincts are in reviewing law society decisions. That is not to say that I agree with the result in every case here. In several cases I disagree strongly with the law society’s decision (e.g., Groia v LSUC), so I’m unlikely to agree with the outcome of the court decision that upholds it. But disagreement with the result does not suggest that the Court was wrong to defer. My observation is that, generally speaking, these cases show courts understanding what their role ought to be, and fulfilling it.
Generally speaking, courts focus on law society reasons, not the substantive issue in the case. Of the 42 post-Dunsmuir judgments, I assessed 75% as focused on the law society’s reasons and analysis, even where the court overturned that decision. In those cases courts did not seem interested in making their own decision and then weighing the law society’s against it. They looked instead at what the law society did, and whether it could be defended as reasonable. And where courts were not reasons-focused, that lack of focus was often understandable. In Trinity Western University v LSBC (2016), for example, the process and substance of the Law Society’s decision was fundamentally flawed such that the Court’s independent assessment of the legal and constitutional issues was understandable (even inevitable). Similarly, in Merchant v LSS (2014), the Court engaged in a detailed review of the issues, but it did so in significant part because of the intensity and detail of the arguments made by the lawyer challenging the decision – there was really no way for the Court to both defer and respond to the lawyer’s case. Certainly sometimes the courts focused more on the substance of the issue before the court than on the law society’s reasoning. In Law Society of Newfoundland and Labrador v Regular (2011) for example, while I am persuaded by the Court’s reasoning, I do not think it acted deferentially in substituting its reasoning about conflicts of interest for the Law Society’s, especially given it reached the same result in the end. Overall, however, the cases did not bear out the concern that Dunsmuir encourages courts to focus on the justifiability of the outcome rather than on the administrative decision-maker’s reasons.
Further, and this is true of both pre- and post-Dunsmuir cases, most of the time I think courts interfered with law society decisions to about the right extent, substantively speaking. They did not tend to second-guess law society decisions about whether a lawyer has committed misconduct, or about the appropriateness of the penalty. They recognized that the law society was in the best position to decide those matters, and they let the law society’s decision and assessment stand. At the same time, however, courts appropriately checked law society over-reach. The Courts overturned decisions where the law society:
- Failed to take into account or even consider a lawyer’s exculpatory explanation for misconduct in assessing the appropriate penalty, or failed to consider joint submissions from the law society and the lawyer on penalty (Guttman v LSM (2010); Hamilton v LSBC (2006); McLean v LSS (2012); Rault v LSS (2009));
- Improperly assessed a delay as prejudicial for only one charge, even though the witness now unavailable was material to both charges (Stinchcombe v LSA (2002));
- Refused to compensate a client for funds paid to a lawyer on the basis that a lawyer who was suspended did not receive money in his capacity as a “lawyer” (Singh v LSA (2000));
- Applied sentencing guidelines that had not been in force at the time of the lawyer’s hearing (Phillon v LSA (1999));
- Had a review panel which did not apply the level of deference review panels of that law society are supposed to apply (LSUC v Abbott (2017); Vlug v LSBC (2017);
- Did not try to make a correct decision, but only tried to make a reasonable one (TWU v LSBC (2016) – although the Court also took the position that the law society denying accreditation to TWU would be an unreasonable violation of s. 2(a), which is more contestable);
- Had a review panel which incorrectly stated that a hearing panel had not made a credibility assessment of a witness, and then proceeded to assess the witness’s credibility without having heard the witness (Mohan v LSBC (2013));
- Imposed a condition that a lawyer provide a psychological assessment of her fitness to practice, when her competence was not at issue (Ritchot v LSM (2010));
- Charged the lawyer with having committed an act intentionally, but then convicted the lawyer on the basis of negligence (Merchant v LSS (2009)).
Of course one does have to remember that it is the court who gets to tell the story in cases like this, with the result that it is perhaps unsurprising that the court tends to look like it is doing the job well. At the same time, however, I chose law society decisions because I am somewhat less likely to be fooled by persuasive judicial writing in this area than I would be in, say, environmental law, where I know nothing. And based on such expertise as I have, it seems to me that the courts mostly have it right in terms of their willingness to interfere with law society decisions. This is the case even where I do not agree with the underlying law society decision such as Histed v LSM (2006) or Groia v LSUC (2016) (with the caveat that the dissent in Groia makes a persuasive case for correctness review and overturning the Law Society’s decision). But the point is that whether a law society reached the same decision I would have on the evidence is not a basis for judicial review, and the courts understand that.
What broader insights follow from this analysis? None, except cautiously and with significant caveats. Law societies have features that set them apart from other administrative decision-makers: the decisions reviewed are quasi-adjudicative; they have extensive procedural protections for the lawyer; many (most?) lawyers are represented by counsel at the law society hearing; most law societies have statutory rights of appeal that take them to appellate courts for judicial review in the first instance, which increases the likelihood of informed judges; the decision-makers are mainly lawyers who are less likely to make bad legal errors; their policy choices rarely come before the court on judicial review. These factors may enhance the likelihood that a court will do judicial review of law society decisions better than they do judicial review generally. On the other hand, courts have concurrent jurisdiction over lawyer conduct through the inherent jurisdiction of the court, and can reasonably be viewed as equally expert to law societies on what constitutes appropriate lawyer conduct. That could have made the court more willing to interfere – no matter how expert the law society, the court would see itself as just as expert. Certainly it was not obvious to me before I read these cases that I was going to find what I did – I expected courts to be much more willing to interfere in law society decisions than they turned out to be.
With that caution and those caveats in mind, I would offer up two tentative conclusions from this review. One is that, as noted before, the effect of Dunsmuir is largely positive or, at worst, not negative. The other, however, is to suggest that the Supreme Court might want to calm down on its standard of review jurisprudence. These cases suggest that appellate courts understand what they are supposed to be doing in judicial review, and they are doing it. Constant tinkering with judicial review by the Supreme Court does not seem to be necessary, and risks unsettling the good work the lower courts, at least in these cases, are doing.