Reasons and Reasonableness in Administrative Law: Delta Air Lines Inc. v. Lukács, 2018 SCC 2

In the Supreme Court of Canada’s latest administrative law decision — Delta Air Lines Inc. v. Lukács, 2018 SCC 2 — it attempted to give some guidance on the relationship between reasons and reasonableness in judicial review cases. It seems like signal rather than noise. There are two interesting aspects, one on the legitimacy of supplementing administrative decision-makers’ reasons for decision after the fact, another on the application of the ‘no-fettering’ principle in reasonableness review.

Dr Lukács is an advocate for the rights of air passengers. He has regularly made complaints to the Canadian Transportation Agency about various matters over which the Agency has jurisdiction, including the terms and conditions of air travel. In his sights on this occasion was Delta Air Lines’ policy relating to obese passengers, alleged to be discriminatory and contrary to the Air Transportation Regulations. The Agency held that Dr Lukács did not have standing to bring his complaint: not being obese himself, he was not personally aggrieved and thus had no private interest standing; and not challenging the legality or constitutionality of state action, he did not have public interest standing. Dr Lukács successfully appealed to the Federal Court of Appeal and won again, 6-3, at the Court. In both forums, there was no dispute that the standard of review was reasonableness.

No Fettering

The problem with the Agency’s approach was that it had fettered its discretion. Section 37 of the Canada Transportation Act allows the Agency to:

…inquire into, hear and determine a complaint concerning any act, matter or thing prohibited, sanctioned or required to be done under any Act of Parliament that is administered in whole or in part by the Agency.

In addressing the admissibility of Dr Lukács’ complaint, the Agency adopted the judicial test for public interest standing, a test that focuses (unsurprisingly) on challenges to the constitutionality or legality of state action (Canada (Attorney General) v. Downtown Eastside Sex Workers United Against Violence Society, [2012] 2 S.C.R. 524). Writing for the majority, McLachlin CJ explained that this is inapt in the context of a complaint brought by a member of the public against a private body, because questions of constitutionality or legality will rarely if ever arise in this scenario: “the Agency suggests the availability of public interest standing to bring a complaint of this type and then, in the same breath, precludes any possibility of granting it” (at para. 17). Adopting the judicial test led the Agency to fetter a broadly drawn discretionary power to to deal with complaints against service providers: it was inconsistent with s. 37.

Moreover, having adopted the judicial test for public interest standing, the Agency proceeded to misapply it. A key component of the Downtown Eastside decision was the insistence on courts taking a flexible, discretionary approach to standing. On the Agency’s approach, however, public interest standing would be impossible to establish: “Downtown Eastside makes clear that at least some plaintiffs will be granted public interest standing. The Agency’s decision, in contrast, allows no complainants to have public interest standing” (at para. 18, emphasis original).

This led to a further problem of inconsistency with the Act as a whole, given the Agency’s sweeping mandate and broad remedial powers:

Applying these tests in the way the Agency did would preclude any public interest group or representative group from ever having standing before the Agency, regardless of the content of its complaint. A complaint by the Council of Canadians with Disabilities…would not be heard. In effect, only a person who is herself targeted by the impugned policy could bring a complaint (at para. 19). 

In dissent, Abella J emphasised that the Agency could legitimately choose to adopt the judicial test for public interest standing (at para. 60). Indeed, it was quite entitled, in her view, to take a narrow of view of the class of potential complainants: “Access to justice demands that courts and tribunals be encouraged to, not restrained from, developing screening methods to ensure that access to justice will be available to those who need it most in a timely way” (at para. 62). Here, Dr Lukács had made a bare complaint that was designed to “engage the Agency in a fishing expedition” (at para. 63).

On balance, I think the majority has the better of the argument here. Even if one accepts that a decision-maker can, in the exercise of a discretionary power, adopt an inflexible rule (see Perry, “The Flexibility Rule in Administrative Law“), the rule must be consistent with the statutory provision conferring the discretion and the statute as a whole. Indeed, the reason that the no-fettering principle tends to be linked to legislative intent is that an inflexible rule will typically (though not inevitably) be difficult to reconcile with statutory text, purpose and context. This case is a useful example, for there were good reasons for concluding that the adoption of an inflexible public interest standing test was inconsistent both with s. 37 and with the scheme of the Act generally.

As to Abella J’s concerns about the value of the underlying complaint, it is worth noting that the majority more or less invited the Agency to refuse to entertain the complaint on some ground other than standing: “there are numerous other ways that the Agency could exercise its discretion under s. 37 of the Act, including examining whether the complaint is in good faith, timely, vexatious, duplicative, or in line with the Agency’s workload and prioritization of cases” (at para. 31). Hint, hint.

Reasons and Reasonableness

The extent to which administrative decision-makers’ reasons for decision can be supplemented after the fact, either by administrative decision-makers themselves (for instance, in affidavits or arguments submitted to the reviewing court) or by judges (relying on a variety of sources including their own imaginations) has been a source of controversy in Canadian administrative law since the Court endorsed the idea that deference should be paid to reasons that “could” be offered in support of an administrative decision (Newfoundland and Labrador Nurses’ Union v. Newfoundland and Labrador (Treasury Board), [2011] 3 S.C.R. 708, at para. 12. See also Dunsmuir at para. 48).

This unsatisfactory state of affairs creates a series of problems. It is unfair to individuals, who argue their cases before administrative decision-makers without knowing the decision-makers’ position on key issues of interpretation and policy (for these can be added later, if the individual is unsatisfied and seeks judicial review). It is inimical to good administration, both because a decision-maker might be tempted to take a less rigorous approach to decision-making if it knows it can supplement its reasons later on, and because a court might impute to an administrative decision-maker a position that would frustrate the efficient and effective attainment of its statutory objectives. It is difficult to reconcile with the legislative choice to assign authority over an area of regulatory policy to an administrative decision-maker, because when a court is allowed to impute additional reasons to an administrative decision-maker it arrogates to judges the power to make judgement calls on issues central to the administrative decision-maker’s statutory mandate. And it may violate a separation of powers between the judicial and executive branch by causing the judges to trespass into matters of administration. See further, with reference to examples, “The Scope and Meaning of Reasonableness Review“, at pp. 814-818.

Despite the concerns I have outlined, the trend since Newfoundland Nurses has been deeply worrisome, culminating in the “quite bizarre” decision in Edmonton (City) v. Edmonton East (Capilano) Shopping Centres Ltd., 2016 SCC 47 where the absence of any reasons for decision did not prevent a majority of the Supreme Court from developing its own, reasonable, interpretation of the relevant statutory provisions. The Federal Court of Appeal subsequently described the effect of Edmonton East as follows:

for a decision to be upheld as being reasonable, it may not even be necessary for the decision-maker to have provided any reasons at all if the record allows the reviewing court to discern how and why the decision was reached and the decision-maker’s conclusion is defensible in light of the facts and applicable law (Canada (Minister of Transport) v. Canadian Union of Public Employees and Sunwing Airlines Inc.2017 FCA 164 at para. 32. Though see the spirited disagreement in 2251723 Ontario Inc. (VMedia) v. Rogers Media Inc., 2017 FCA 186, at paras. 26 and 32-33).

In Delta Air Lines, McLachlin CJ at least tried to arrest this trend, distinguishing Newfoundland Nurses and Alberta (Information and Privacy Commissioner) v. Alberta Teachers’ Association, [2011] 3 S.C.R. 654 as cases involving “non-existent or insufficient” reasons where supplementation “may be appropriate”, whereas here “the Agency provided detailed reasons that enumerated and then strictly applied a test unsupported by the statutory scheme” (at para. 23).

McLachlin CJ referred only to Alberta Teachers’ and not Edmonton East — passed over in silence like an unloved distant relative — but these cases have an important shared feature: in both, a failure to supplement reasons after the fact would have undermined the deference otherwise owed to the decision-maker. Accordingly, it seems that the signal McLachlin CJ is trying to send is that after-the-fact supplementation is appropriate only where a failure to do so would prevent a reviewing court from according due deference to an administrative decision-maker. This would significantly restrict the ambit of the idea that courts should defer to reasons that could have been offered in support of a decision.

Of course, there is nothing particularly objectionable about Newfoundland Nurses itself, if the decision is understood to stand for the proposition that administrative decisions should be read charitably, with courts giving administrative decision-makers the benefit of the doubt where reasons are sparse or not perfectly satisfactory, and not for the proposition that reasons can be rewritten in order to bolster the underlying decision. This also forms part of the signal McLachlin CJ is trying to send: “while a reviewing court may supplement the reasons given in support of an administrative decision, it cannot ignore or replace the reasons actually provided” (at para. 24). To do otherwise would “undermine, if not negate, the vital role of reasons in administrative law” (at para. 27). If this signal is received by lower courts and administrative decision-makers, Canadian administrative law will be the better for it.

McLachlin CJ explicitly endorsed an important passage, penned by Rennie J (as he then was) in Komolafe v. Canada (Citizenship and Immigration), 2013 FC 431, at para. 11:

Newfoundland Nurses is not an open invitation to the Court to provide reasons that were not given, nor is it licence to guess what findings might have been made or to speculate as to what the tribunal might have been thinking.  This is particularly so where the reasons are silent on a critical issue.  It is ironic that Newfoundland Nurses, a case which at its core is about deference and standard of review, is urged as authority for the supervisory court to do the task that the decision maker did not do, to supply the reasons that might have been given and make findings of fact that were not made.  This is to turn the jurisprudence on its head.  Newfoundland Nurses allows reviewing courts to connect the dots on the page where the lines, and the direction they are headed, may be readily drawn.  Here, there were no dots on the page.

In this case, there was no justification for supplementing the decision-maker’s reasons. Doing so would cause the reviewing court to supplant the primary decision-maker, the Agency. For the majority, this was precisely the sin committed by Abella J in dissent, who raised features of the complaint that were not relied upon by the Agency.

Another salutary feature of McLachlin CJ’s approach is her willingness to remit the decision to the Agency. Underpinning the worrisome post-Newfoundland Nurses trend — sometimes explicitly, sometimes implicitly — is a concern for efficiency and cost-effective administration. Why remit a matter to an administrative decision-maker for costly, time-consuming further procedures where it has become obvious, over the course of the litigation, that the ultimate result would be that the same decision would be reached, supported by more convincing reasons? In my view, this does not outweigh the important concerns outlined above. And McLachlin CJ was very firm: “The better approach is to send this matter back to the Agency for reconsideration in its entirety” (at para. 30). Indeed, her justification for doing so — “the legislator’s intention to give deference to the Agency’s determination of its complaints process” (at para. 30) — can be generalised to most administrative decision-makers. They, not the courts, are the primary decision-makers. With this in mind, reviewing courts should be much more willing to remit matters even when tempted to resolve them authoritatively. And if a decision-maker is put to extra time and expense, so be it, pour encourager les autres.

Finally, although Delta Air Lines is, on the whole, a laudable decision, one unfortunate implication of all this is that the administrative decision-maker that gives no reasons will be in a better position — by virtue of its ability to provide additional reasons if judicial review proceedings are brought — than an administrative decision-maker that gives sparse reasons. There might therefore be a perverse incentive not to provide reasons at all. However, given the wide scope of the common law and statutory duties to provide reasons — the giving of which is, moreover, facilitated by modern technology — administrative decision-makers would be asking for trouble were they to avoid giving reasons for decisions. It will be difficult for ‘bad‘ decision-makers to game the relationship between reasons and reasonableness.


This content has been updated on January 22, 2018 at 16:25.