Consistency in Administrative Adjudication: Canadian Association of Refugee Lawyers v. Canada (Citizenship and Immigration), 2019 FC 1126 and Shuttleworth v. Ontario (Safety, Licensing Appeals and Standards Tribunals), 2019 ONCA 518

In two important recent decisions, Canadian courts have had to consider the lawfulness of internal administrative arrangements designed to promote consistent decision-making. On both occasions, the arrangements ran afoul of the principles of administrative law. In Shuttleworth v. Ontario (Safety, Licensing Appeals and Standards Tribunals), 2019 ONCA 518 a process of peer review of draft decisions was held to be unlawful and, in Canadian Association of Refugee Lawyers v. Canada (Citizenship and Immigration), 2019 FC 1126, a jurisprudential guide issued by the Chairperson to members of the Immigration and Refugee Board met a similar fate. Shuttleworth related to the definition of “catastrophic impairment” in Ontario’s accident benefits scheme, whilst CARL concerned decision-making by the Refugee Appeal Division in respect of refugee claimants from China, India, Nigeria and Pakistan.

Both cases were decided against the backdrop of a delicate balancing act. On the one hand, the role of peer review and the jurisprudential guide is to enhance the accuracy and consistency of decision-making. Achieving accuracy and consistency can be a challenge for multi-member tribunals, especially those whose members are dispersed across a province or the country. The spectre of inconsistent decision-making is never too far away from such bodies: “Draw one adjudicator and one interpretation will be applied; draw another and the opposite interpretation will be applied. Under the rule of law, the meaning of a law should not differ according to the identity of the decision-maker” (Wilson v Atomic Energy Agency of Canada [2015] 4 FCR 467, at para. 52 per Stratas JA). This is a particular concern in immigration decision-making, where different conclusions might be reached in respect of individuals who come from the same geographic region or who have had similar political affiliations in the past (factors which may bear on applications for status in Canada). And in terms of accuracy, it might be wise on difficult questions of law or policy (such as the definition of “catastrophic entitlement” at issue in Shuttleworth) for decision-makers to consult with colleagues with a view to reaching the the best possible decision.

On the other hand, s/he who hears must decide; tribunal decision-makers are the ones charged with deciding cases, not their superiors, and they must do so “in accordance with [their] own conscience and opinions” (IWA v. Consolidated‑Bathurst Packaging Ltd., [1990] 1 S.C.R. 282 at 332), unfettered by (for example) soft law instruments seeking to confine their discretion.

Accordingly, a balance must be struck between deciding individual cases on their merits, thereby treating the individuals concerned fairly and respecting the adjudicative independence of the decision-makers, and the interests of accurate and consistent decision-making. As Crampton CJ explained in CARL:

…in the administrative law context, it can be entirely appropriate to embrace tools such as guidelines to influence, in a general way, the manner in which decisions are reached. In this regard, a legitimate type of general influence can include identifying factors, sources of information, and even particular information that can be helpful to consider. Indeed, I consider it to be permissible to go further and encourage such information to be taken into account, so long as it is made clear that decision-makers remain completely free to reach their own conclusions, based on the facts of each particular case (at para. 93).

The particular problem in Shuttleworth was that decisions were “generally sent…to the executive chair without assent or input from the adjudicators” and no written policies were in place permitting adjudicators to refuse to participate in the peer review process (at para. 18, recording a factual finding of the Divisional Court). This gave rise to a reasonable apprehension that adjudicators were compelled to consult (at para. 33). As the Supreme Court has noted, however, a “consultation proceeding could not be imposed by a superior level of authority within the administrative hierarchy, but could be requested only by the adjudicators themselves” (Ellis-Don Ltd. v. Ontario (Labour Relations Board), [2001] 1 SCR 221, at para. 29). Comments here were received “on whether the correct legal test and jurisprudence has been identified and applied” (at para. 31), including from the executive chair, who recommends whether adjudicators should be reappointed (at para. 32). Ultimately, “by deciding to provide comments on the decision without seeking the adjudicator’s permission, the executive chair imposed the review on the adjudicator” (at para. 50). Compulsion was central to Hourigan JA’s analysis, as in the following key passage:

The absence of a formal policy protecting the adjudicator’s right to decline to participate was significant in an environment where the procedure manual made no reference to the voluntariness of review by the executive chair and [staff counsel] Mr. Cowan’s own evidence was that adjudicators were expected to participate in the review process. While Mr. Cowan’s evidence was that there was no means to compel adjudicators to participate, he did not give evidence that SLATSO communicated to adjudicators that they were free not to have their drafts reviewed by the executive chair. The principal safeguard the appellants point to is the executive chair’s inability to lawfully compel the adjudicator to change her mind. However, Consolidated-Bathurst establishes that this is not a sufficient safeguard: at p. 333 (at para. 39).

Hourigan JA’s analysis is persuasive, given the unusual facts of the instant case, where a whistleblower had pierced the veil of deliberative secrecy sufficiently to suggest that the peer review had resulted in a material change to the draft decision (see further at paras. 6-13). However, even a slightly better-drafted peer-review policy would presumably have led to a different result. If the policy had emphasized the adjudicator’s ability to refuse to participate in the peer-review process and perhaps added that the executive chair is expected to review draft decisions bearing on major questions of policy, it would be difficult to find it unlawful. The problem of the executive chair’s reappointment power would persist, of course (and will continue to persist as long as Canadian courts (outside Quebec) respect adjudicative independence more in the breach than the observance), but otherwise the adjudicator’s ability to decide cases on their merits would be preserved.

Pressure from above was also a concern for Crampton CJ in CARL. Even though there was “no evidence that Board members would face sanctions or other adverse consequences for not applying any of the impugned JGs, common experience would suggest that at least some Board members would feel pressured by such repeated statements of expectation from their superiors regarding how they should conduct themselves” (at para. 134). The problem was that “at least some [IRB] members would feel pressured to adopt the factual determinations in the JG” (at para. 137), because the policies at issue were not expressed to be non-binding (at para. 139).

Again, however, relatively minor tweaks would permit the Board to achieve the same result: “it would be entirely legitimate for the Chairperson to encourage Board members to take the JG into account in cases with similar facts” and the Chairperson could even encourage them to follow the guide “so long as it was also made very clear that they are completely free to depart from the JG based on the particular facts of the case before them” (at para. 142).

Indeed, depending on one’s view of the role of stare decisis in administrative decision-making structures, the policies at issue in CARL were not especially problematic. Crampton CJ considered that the policies had a “distinct mandatory aspect”:

…because of the statement that “RPD and RAD members are expected to apply JGs in cases with similar facts or provide reasoned justifications for not doing so.” Given that statement, Board members would be subject to an explicit expectation to adopt the above-mentioned findings, unless they were prepared to provide reasoned justifications for failing to do so. In cases with similar facts, it is reasonable to expect that some Board members who might be unable or unwilling to provide such justifications may very well feel pressured to adopt the factual findings in question because of the instruction that this is what Board members are expected to do. This is particularly so in light of the fact that the statement of expectation has been conveyed to the Board’s members repeatedly, including in its Policy on the Use of Jurisprudential Guides, an e-mail from the Chairperson dated 21 July 2017, and an e-mail of the same date from the Deputy Chairperson (RPD) (at para. 133).

My view, as set out in “Stare Decisis in Canadian Administrative Law” (2016), is to the effect that administrative decision-makers within the same structure may legitimately differ amongst themselves, as long as they provide a reasoned justification for a refusal to follow a previous decision:

In general…there is a “strong case for branding as reviewable those cases where statutory authorities inexplicably fail to act consistently”. It follows that where a decision-maker departs from a previous decision, the departure must generally be accompanied by an explanation justifying the departure; the previous decision provides a “direct contextual comparison against which” the reasonableness of the new decision can be assessed. Once this criterion is met, however, the departure will be upheld as a reasonable decision.

(2016) 49(1) Revue juridique Thémis 757. See also Céré v. Canada (Attorney General), 2019 FC 221

On that approach, there is nothing particularly objectionable about requiring decision-makers, especially in a multi-member tribunal, to justify departures from starting points established by the tribunal’s jurisprudence. Of course, as Crampton CJ notes (at para. 138), the ability to set starting points must be exercised with caution, but as long as these starting points are reasonable (and subject to review for reasonableness) the goal of achieving greater consistency can be achieved without fettering the discretion of individual members.

Finally, as a result of the continuing uncertainty about the standard of review in Canadian administrative law, it is not clear which analytical framework reviewing courts should apply to cases such as these. In both of these cases, substantive policies were at issue, relating to the circulation of draft decisions in Shuttleworth and the appropriate response to given factual situations in CARL. In Shuttleworth, Hourigan JA approached the matter as one of bias, asking whether a reasonable and well-informed observer “would think that it is more likely than not that the decision-maker would decide fairly” (at para. 23). In CARL, meanwhile, Crampton CJ viewed the central question as whether individual decision-makers’ discretion was fettered by the jurisprudential guide. As “the fettering of a decision-maker’s discretion is per se unreasonable”, it was unnecessary for him to identify the standard of review (at para. 57). There is at least an argument, however, that the policies at issue in these cases should have been subject to reasonableness review. For where an administrative decision-maker adopts a policy under its parent legislation to give effect to its statutory objectives, the standard of review is typically reasonableness. This is not a criticism of Crampton CJ and Hourigan JA, who were able to cite authorities to justify their approaches. These authorities are, however, in tension with the Supreme Court’s contemporary approach to substantive review (see here and here). Unfortunately, this tension did not feature in the Supreme Court’s administrative law “trilogy” and is unlikely to be resolved any time soon.

This content has been updated on September 30, 2019 at 15:35.