A Pepa Talk on Reasonableness Review and Remedies: Pepa v. Canada (Citizenship and Immigration), 2025 SCC 21

Subsection 63(2) of the Immigration and Refugee Protection Act, SC 2001, c 27, provides: “A foreign national who holds a permanent resident visa may appeal to the Immigration Appeal Division against a decision to make a removal order against them made under subsection 44(2) or made at an admissibility hearing”.

The appellant in Pepa v. Canada (Citizenship and Immigration), 2025 SCC 21 held a visa when she entered Canada. But on her entry into Canada, she was referred for an admissibility hearing to determine her entitlement to remain in the country.

By the time the hearing at the Immigration Division rolled around, her visa had expired. She received a negative decision in the form of a removal order.

She sought to appeal to the Immigration Appeal Division. Whereupon the IAD determined that it did not have jurisdiction, on the basis that precedent within the IAD and from the Federal Court and Federal Court of Appeal made clear that a prospective appellant must hold a valid visa at the time of the appeal.

Both the Federal Court and Federal Court of Appeal found that the IAD’s interpretation of s. 63(2) was reasonable. A majority of the Supreme Court thought differently, though.

Martin J’s key proposition was that it is unreasonable for an administrative decision-maker “to rely on clearly inapplicable or distinguishable case law — like cases in different areas of the law or cases addressing different statutory provisions — without justification and explanation of its continued relevance to the matter at hand” (at para. 66). That was the error the Division had slipped into:

In this case, the precedents the IAD relied on were not sufficient to resolve the statutory interpretation question before it, nor did the IAD justify or explain their continued currency where they concerned an outdated statutory provision or starkly different facts. As such, it was unreasonable for the IAD to constrain itself to their holdings (at para. 68).

…

Reliance on a clearly distinguishable non-binding case, and the subsequent IAD decisions which followed it, without further analysis themselves, cannot be reasonable without explanation of the reasoning behind such a conclusion. No explanation is present in the IAD’s reasons here (at para. 76).

One case involved a predecessor provision (at paras. 69-72). But as Martin J commented, “[t]he IAD used Hundal as binding authority when its jurisprudential force could not simply be assumed but should have been explored, explained, and justified” (at para. 72).

A series of IAD cases ultimately relied on the Hundal case as well. Here, the problem was that “because Hundal was used as the basis for the other IAD decisions considered, it means the reasons contain a near total absence of any attention to the text, context, and purpose of the very legislative provision over which Ms. Pepa and the Minister asserted competing interpretations” (at para. 75).

Another case was about a revoked visa, not an expired visa, and accordingly the observations about expired visas were obiter (at paras. 77-78). This was also true of yet another Federal Court case (at paras. 80-82).

And, believe it or not, there was also a case interpreting the predecessor provision in a manner favourable to the appellant — but it was “brushed aside” by the IAD, whereas an unfavourable precedent was waved through (at para. 83).

Ultimately, Martin J concluded, the precedent relied upon by the IAD was not dispositive:

Overall, the FCA below erred in concluding that based on the IAD’s review of the case law, it was reasonable for the IAD not to undertake its own statutory interpretation analysis. While Vavilovstates that precedent will act as a constraint on what the decision maker can reasonably decide, this only applies to precedent on the issue before it, or precedent on a similar issue. A decision maker needs more than a few citations to cases relying on a different provision, or a clearly distinct factual matrix, to determine the issue. Though failure to conduct a statutory interpretation analysis is not fatal on its own, where the case law available to the decision maker is not sufficiently material or binding, the analysis cannot simply stop without ensuring that due consideration has been given, according to the modern principle of interpretation, to the competing interpretations asserted by the parties (at paras. 84-85).

This decision neatly illustrates some of the points I made in my recent paper on reasonableness review after Vavilov (here, at pp. 30-33).

Most importantly, it is the task of the reviewing court to determine whether a particular constraint is relevant. Here, the relevant constraint was judicial and administrative precedent. Ultimately, the majority concluded that the precedent relied on by the IAD was not relevant to the issue at hand. There was no deference to the decision-maker on the determination of whether the precedent was relevant. The fact that no deference was given on this point might lead one to think that the approach of the majority was not respectful of the autonomy of the IAD. Perhaps Martin J had lunched on the forbidden fruit of disguised correctness review? That is never a good dietary choice for a reviewing court but, in this instance, the majority did not succumb to temptation: it is always the role of the reviewing court to determine whether a particular constraint is relevant.

The fatal flaw here was the assumption on the part of the IAD that the precedent was relevant. That assumption did not become any more justifiable because it had been made in a series of cases. Fundamentally, the IAD had not explained why the precedent was relevant. It had assumed, without comment, explanation or justification that decisions relating to the predecessor provision and/or visa revocation cases were also relevant to the contemporary provision and visa expiration cases. However, as Martin J pointed out, this is not self evident. The relevant question was: are there relevant precedents on the issue of whether an expired visa has the same consequences as a revoked visa for the purposes of s. 63(2)? The decision-maker never asked itself that question, and did not explain or justify why it had not asked that question. Accordingly, there was nothing for the majority to defer to.

The dissenting judges had a different view of the relevance of precedent. As Côté and O’Bonsawin JJ remarked in respect of the statutory change, “the distinctions between these two provisions do not affect the reasonableness of the IAD’s decision” (at para. 194). And they also rejected the expiration/revocation dichotomy of the majority: “While the facts of that particular case [Ismail] deal with revocation, when read as a whole, it is clear the reasons of de Montigny J. are animated by the validity of the visa itself” (at para. 204). However, on these points, I do not think they deferred to the reasons of the decision-maker. This was a threshold question about relevance which the court ultimately had to answer for itself.

In my view, there is no interference with the autonomy of the decision-maker in such a circumstance. Under Vavilov, deference can only be given on the basis of reasons. If no reasons have been provided, no deference is possible.

Another difficult question raised by this case is what happens next. Having established that the decision was unreasonable because the decision-maker had relied on precedent that, properly read, did not govern the situation at hand, the majority went on to engage in an exegesis of the statutory scheme. They ultimately concluded, having reviewed the text, purpose, and context of the relevant statutory provisions that there was only one reasonable answer to the interpretation of s. 63(2). I do not intend to provide chapter and verse on this, but I will highlight two interesting aspects. For one thing, Martin J relied on soft law produced by a government department to support her interpretation of s. 63(2) (at paras. 110-112): this is the first instance of which I am aware of the Supreme Court relying on soft law to interpret the meaning of a statutory provision in a context not involving the exercise of discretion (Agraira, Kanthasamy and Dow Chemical). For another, she relied heavily on the ‘harsh consequences’ constraint from Vavilov:

In my view, the IAD did not give sufficient consideration to the relatively significant consequences of the decision for Ms. Pepa. Though the stakes here are not as high as in the penal context, the consequences are nonetheless severe. Further, the IAD’s failure to address key factors of statutory interpretation at all in its reasons shows it did not explain why its decision respects Parliament’s intention, let alone “best reflects” Parliament’s intention. Parliament intended an efficacious appeal process, and the IAD’s reading makes this process all but illusory in cases where the visa has expired before the removal order is issued. The reasons ought to have demonstrated that the decision maker considered the consequences of the decision and whether such harsh personal consequences were justified in light of the facts, the law and Parliament’s intention. This did not happen here (at para. 119).

I do not have any strong views on whether the outcome of the analysis was good, bad or indifferent. However, I think the majority was ill-advised to engage in this exercise at all. As Rowe J explained in his concurring reasons:

In this, I am mindful of the uncontroverted submission by the Minister of Citizenship and Immigration that the IRPA and the Immigration and Refugee Protection Regulations, SOR/2002-227, contain multiple provisions that refer to “holding” or a “holder” of a visa (or other document) (R.F., at paras. 86 and 92). The Minister of Citizenship and Immigration submits that adopting Dorinela Pepa’s interpretation of the term “holds” in the context of s. 63(2) of the IRPA would lead to “absurd” consequences when applied to other provisions (para. 92). Ms. Pepa responds that the interpretation she prefers could be limited to the meaning of “hold” in circumstances similar to those in this case.

If this Court itself interprets the relevant provision, there may well be consequences for the legislative scheme which we cannot contemplate, but that the IAD would more readily appreciate. This very much favours referring the matter back with guidance rather than deciding it on the basis of a single reasonable interpretation (at para. 150-152).

It was inevitable that the majority would encroach on the autonomy of the decision maker by engaging in the statutory interpretation exercise that the decision-maker would ordinarily have engaged in. Properly directed on the issue of precedent, the IAD could have done its own analysis. Had it done so, it may have arrived at a different conclusion given its knowledge of the statutory scheme. In all events, here the majority was engaging in an exercise that fell within the domain of the decision-maker. Rowe J worried that this was “disguised correctness review” (at para. 148) but there was really nothing disguised about it at all. To return irreverently to my Garden of Eden metaphor, the majority reasons did not even sport a fig leaf: it was stark-naked correctness review on a point that would have been better left to the IAD.

What lessons does this Pepa-talk hold for administrative tribunals and reviewing courts? As far as administrative tribunals are concerned, this should serve as a useful reminder that careful attention to the underlying justification for, and scope of, internal and judicial precedent is extremely important. But I do not think anything here should be taken to undermine the importance of consistency in adjudication. To put it bluntly, it is often better to be consistently wrong than consistently inconsistent. In doing so, however, it is important to return to the source and analyze, even if briefly, the underlying justification and scope of any precedent relied upon, especially where there are indicators — legislative change, harsh consequences or inconsistency with soft law — that the precedent might not be sound.

For reviewing courts, this admonition of Martin J should not be overlooked:

Reviewing courts must exercise caution not to overstep their role when examining the soundness of the precedents that were relied on. The role of a reviewing court in such an instance is to ensure that the duty of justification has been discharged, not to wrestle with the correctness of the past administrative precedents, nor to overturn by proxy a judicial precedent (at para. 68).

As far as refusing to remit a matter is concerned, I am inclined to think that this case was somewhat unusual. But it is becoming harder and harder to ignore the fact that when the Supreme Court finds an administrative interpretation of law to be unreasonable, it also finds that there was only one possible, acceptable interpretation of the provision that should be carved in stone. We are coming quite close to receiving a ‘signal‘ in that regard, if indeed the signal has not already been sent but some of us just refuse to see.

This content has been updated on June 27, 2025 at 22:57.