Doré or Door B

by on

An unpopular opinion: Doré [2012 SCC 12] is fine. In the 8 years since it, we haven’t seen a downfall of civilization and the Rule of Law (well, not in Canada at any rate).

The core of Doré is the idea that when an administrative decision maker (ADM) makes a decision that implicates Charter rights, the ADM’s decision is owed deference: Doré at para. 57. The question was whether this deference should be assessed according to the Oakes test, or if it should be assessed according to administrative law principles. Doré decided that it should be assessed according to administrative law principles.

One implication, for those who inveigh against Doré, is that administrative actions that would be unconstitutional if assessed by Oakes would be constitutional if assessed by Doré. This would appear to flout the principle that the legislatures cannot provide to a delegate power that the legislature itself does not possess. Moreover, where the government bears the burden of justifying an infringement under Oakes, the applicant for judicial review bears the burden in reasonableness review. Doré is also criticized for potentially reversing that burden.1Law Society of British Columbia v. Trinity Western University, 2018 SCC 32 (per Côté and Brown JJ) at paras. 312-314.

I would approach this question rather differently.

I would separate the relevant decisions into two: (1) the constitutionality of the provision in the legislation that empowers the ADM to make decisions; (2) the accordance of the ADM’s decision with the power delegated to them.

The first question of constitutionality relates to whether the empowering provision complies with the Charter when it delegates the power to decide Charter questions. Whether it does or not depends on how these decisions are reviewed by the courts. There are three options: correctness review, reasonableness review, and no review.

There is clearly no constitutional problem if the ADM’s decision is subject to correctness review. On correctness review, the ADM’s decision may come first, but any unconstitutional decision by an ADM can be challenged in the court. Given the general view that “what is constitutional” is the same as “what courts decide is constitutional”, there is no risk of this legislation leading to unconstitutional actions by a public authority.

But what if the ADM’s decision is not subject to correctness review, as is the case under Doré? This should create the possibility that some decisions will be made that will be upheld by a court on judicial review, but if considered at first instance by a Court would be struck down. As I understand it, it is the gap that Doré creates between what courts would uphold as constitutional and what ADM’s will uphold that causes part of the consternation. Holding to the same logic that “what is constitutional” is the same as “what courts decide is constitutional”, the empowering provision would essentially empower an ADM to violate the Charter. Taking the logic of the anti-Doré crowd to its conclusion, if Doré is good law, then the empowering provisions should be unconstitutional.

As I see it, there are two competing theories in favour of Doré.

The first is an argument from expertise. On this view, ADMs may be relatively more expert than Courts both in matters of law (in determining the scope and value of rights) and in matters of fact (e.g., in appreciating the justifications for an infringement). This theory is propounded in Doré itself 2at paras 30, 35, 47, 48.

The second is an argument from efficiency. On this view, ADMs exist to make decisions efficiently, even if they sometimes do so incorrectly. Governments are seen to provide decision-making power to ADMs rather than courts because the ADMs enjoy a decision cost advantage (the cost of deciding) that exceeds their error cost disadvantage (the cost of getting a decision wrong). Some ADM decisions will be wrong, but it may be better to get all decisions made relatively quickly and relatively cheaply than to incur the additional cost needed to confirm that the decision is correct. This theory is also nodded at, albeit in not precisely these terms, in Doré3at para. 51.

Either of these arguments could be deployed to support Doré against the above critique. The logic would be that the infringement of rights occasioned by empowering applying Doré is itself a “reasonable limit[] prescribed by law” that “can be demonstrably justified in a free and democratic society”, with the justification being either expertise or efficiency. Context matters for evaluating whether either of these arguments is convincing.

The expertise argument faces significant challenges in providing sufficient justification in modern Canada. The primary challenge with this argument is that it puts the proverbial foxes in charge of guarding the henhouse: it puts some public servants, who cannot possess the independence of the courts, in charge of deciding what other public servants can do. Ostensible greater understanding and appreciation of the importance of government objectives can simply be a predisposition to believe the government’s case for an infringement.

The examples given in Doré of how ADMs have “by virtue of expertise and specialization, particular familiarity with the competing considerations at play in weighing Charter values”4at para 47, examples following at paras. 48-50 support that ADMs have familiarity with only one of the competing considerations: the statutory objectives that justify infringement. It is unclear from Doré why ADMs would also have superior ability to understand the scope of rights or the appropriate balance between an individual’s rights and the infringement.

The efficiency argument is much more compelling in modern Canada. Trials can take years. Courts are a robust process, but they are not a fast process. Could it not be demonstrably justified to give an ADM authority over constitutional issues, if they got 5% of decisions wrong (where “wrong” is “reasonably but incorrectly”), but made all decisions twice as fast?

To be clear, I do not seek to provide a justification to dispense with judicial review altogether (even setting aside s. 96 5Vavilov, 2019 SCC 65 at para. 24): a free and democratic society does not permit one person to be egregiously harmed ‘for the greater good’. Maintaining a role for courts to prevent egregious violations of rights authorized by ADMs addresses this problem. This is the very definition of reasonableness review: a court that does not simply inject its own opinion, but also a court that ensures the decision stayed within reasonable bounds.

I will add two further observations.

First, Doré absolutely gives legislatures powers to do by delegation what they could not do directly. A delegation to a decision-maker constrained only by reasonableness review may itself be reasonably necessary in a free and democratic society. This provides a stronger rationale to justify an infringement under s. 1 than if the same effects were accomplished directly by the legislature. If the legislature acts directly, then only the substantive purposes can justify the infringement; when the legislature acts indirectly, the reason for acting indirectly also can provide some justification for the infringement. This also suggests a possible grounding for the reversed-onus in Doré: the onus remains on the government, but the simple fact of delegation itself provides a preliminary justification, thereby shifting the burden back onto the claimant.6 There is a relatively minor problem here, which is that the government should be the one to advance the justification. But the “common law” of administrative law might generally be seen as a series of legal presumptions about legislators’ factual purposes, and one could see this reversal as simply adding another presumption to that stack.

Second, for the efficiency argument to be compelling, the total decision cost (approximately, costs to parties of litigating + costs to public of deciding) must be significantly lower when an ADM faces reasonableness review than when the same ADM faces correctness review.

The truth of the efficiency argument is less a normative question than it is an empirical one, but the empirics are complicated. The causal pathway for the efficiency argument depends first on how courts apply Doré. If courts apply “disguised correctness” review when they are supposed to apply reasonableness review, the effect of Doré will be minimal. If Doré is, in fact, followed, then the second question is whether ADMs change their practices when subject to reasonableness review rather than correctness review. The third question is whether parties change their practices before ADMs who are subject to reasonableness review. If this efficiency argument has empirical validity, the adoption of reasonableness review by the Courts should occasion a change of practice before ADMs.

Vavilov may have created a pseudo-natural experiment that would help evaluate whether these empirical inferences are valid. It, famously, Housenized the standard of review for administrative decisions subject to a statutory “appeal” clause 7Vavilov at para 37. The Housen standard lies in between correctness and reasonableness. If the efficiency argument is correct, then decisions of the ADMs subject to Housen review should become more costly post-Vavilov than pre-Vavilov. Moreover, the increase in cost should be greater for ADMs subject to a statutory appeal clause than for ADMs that are not so subject. The cost could be measured in various ways, including looking at the length of written decisions, the proportion of written decisions, the proportion of self-represented litigants, and, to the extent possible, the billings of lawyers before the tribunal.

In sum:

  • Yes, Doré leads to administrative decisions being upheld that would not otherwise be upheld
  • Yes, Doré leads to the legislature being able to delegate power that it itself does not possess
  • No, this isn’t inherently a problem
  • Yes, the arguments from non-court expertise are bad