Rift on the Supreme Court Bench? Cont’d (Fraser and G)

In my November 3rd Slaw post on the Supreme Court of Canada’s decision in Fraser, I considered the division on the Court relating to the interpretation of section 15(1) of the Canadian Charter of Rights and Freedoms. The majority decision, written by Abella J., emphasized a broad interpretation, stressing the significance of adverse effects discrimination and the goal of substantive equality. In their dissent, Brown and Rowe JJ. applied a narrower interpretation, as did Côté J. in her separate dissent. Now we have Ontario (Attorney General) v. G, which not only reminds us of the cleft in interpreting section 15, but also extends much the same split in the justices’ approach to the Charter to the remedial issues of a suspended declaration of invalidity and individual exemption.

G involved Christopher’s Law (Sex Offender Registry), 2000, which established the Ontario sex registry, and the different treatment of persons found guilty of a sexual offence and those held to be not criminally responsible on account of mental disorder (“NCRMD”). Both have their personal information added to the provincial registry and must report to the police station annually or as necessary to ensure their information is current; however, persons found guilty may be granted an exemption, be removed from the registry or relieved of reporting requirements on an individual assessment, while this possibility is not available to persons NCRMD even if they are granted an absolute discharge.

G., who was found NCRMD for sexual offences in 2002, was granted an absolute discharge in 2003 and his information was placed on the registry in 2004. He challenged Christopher’s Law under sections 7 and 15(1) of the Charter on the basis that it treated persons NCRMD differently from those found guilty of committing sexual offences. He was not successful in relation to section 7, but he was in relation to section 15(1).

In Fraser, the difference in approaches to section 15 affected how the majority decided and how the dissenting justices would have decided the outcome of the case. The majority held that the RCMP pension regime that treated members taking a full time leave and those who took advantage of a job-sharing program differently was discriminatory against the job-sharers and not justified, while the dissenting judges found no discrimination. In G, however, the court was unanimous that the regime was discriminatory against persons found NCRMD; regardless, the emerging different schools of thought about how to interpret and apply section 15(1) came into focus.

In G, Karakatsanis J., writing for the same judges constituting the majority in Fraser, reviewed the interpretative territory of section 15(1) in paragraphs 39 to 47, stressing the significance of recognizing adverse impact discrimination and substantive equality. Justices Côté and Brown did not consider it necessary to discuss the underlying principles of section 15(1), saying that the section 15(1) issue could be dispensed with “easily”, since the law was facially discriminatory and the distinction in treatment between persons found guilty and those found NCRMD “exacerbates pre‑existing disadvantage by perpetuating the stereotype that persons with mental illness are inherently dangerous”. They criticized Karakatsanis’s section 15 analysis in no uncertain terms:

Our colleague, however, goes further, and in extensive obiter dicta discusses adverse‑effects discrimination and “substantive equality” (paras. 41‑69). Her doctrinal statements are not remotely relevant to the issues raised by this appeal, especially considering this is not an adverse‑effects case. The distinction in this case is facially apparent….

Thus, our silence on paragraphs 41‑69 of our colleague’s reasons should not be taken as tacit approval of their content. We simply do not see them as offering actual reasons for her judgment, but “commentary . . . or exposition” instead… (paras. 223-224)

Justice Rowe simply records his agreement with Côté and Brown JJ., as far as section 15(1) of the Charter is concerned (G, Rowe J., para. 185).

It is worth noting, however, that Karakatsanis J.’s analysis subsequent to paragraph 47 is primarily a response to the Attorney General of Ontario’s arguments and an explanation of why the difference in treatment denies the substantive equality rights of persons with mental disorders. The first part of Karakatsanis J.’s section 15(1) analysis is a reminder of the principles underlying the majority’s interpretation, and particularly reflects the understanding of substantive equality that is important to assessing why even facially discriminatory provisions reflect stereotypes and disadvantages, while much of the rest links those principles to the arguments in the case.

Beyond the continuing need to emphasise the dispute over the interpretation of section 15(1), which had no actual impact on the outcome, there was also disagreement on the court’s remedial jurisdiction, which would have affected the outcome for G, at least.

The Ontario Court of Appeal had held Christopher’s Law to be unconstitutional because of its different treatment of persons found guilty of sexual assault and those found NCRMD, suspended the declaration of invalidity for a year and granted G a constitutional exemption resulting in the removal of his information from the registry.

At the Supreme Court, Karakatsanis J., for the majority, Côté and Brown JJ., in partial dissent and Rowe J. in a concurring judgement (perhaps best described by implication as a partial dissent) all weighed in on when the court should grant a suspension of a declaration of invalidity and when it should grant a constitutional exemption and, more broadly, on whether the 1992 decision in Schachter should continue to guide the determination of the appropriate remedy.

Justice Karakatsanis wrote the majority decision on behalf of the same six judges constituting the majority in Fraser; she suspended the declaration of invalidity and granted G an exemption from Christopher’s Law. Justices Côté and Brown would have granted the suspension, but not granted G an exemption. Justice Rowe agreed with Côté and Brown JJ.’s partial dissent relating to ordering an individual constitutional exemption, although not with their approach to suspending a declaration of invalidity. Of all the judges, he was the only one who declared himself satisfied with the remedial approach in Schachter.

Behind these conclusions, however, the judges sought to address the larger question of judicial remedial jurisdicton and in doing so, continued to exhibit the same dissension evident in Fraser and reinforced even in the discussion of section 15 in G.

Suspending a declaration of invalidity in order to allow the legislature to bring the impugned law into conformity with the constitution began in 1985 with the creative response to the dilemma resulting from the Supreme Court of Canada’s finding that almost the entire Manitoba legal system was unconstitutional because statutes were not in French as well as English as required by section 23 of the Manitoba Act, 1870, part of the Constitution (see Schedule to the Constitution Act, 1982). The court described the situation that would result if they issued a simple declaration: “a legal vacuum will be created with consequent legal chaos in the Province of Manitoba” (Manitoba Language Rights, para. 55). Legal institutions deriving their powers from legislation enacted in English only would be acting without legal authority and all rights relying on legislation would be open to challenge. This would undermine the rule of law, which requires a system of positive laws, and therefore the Court determined that the rule of law required a suspension of the declaration for a specified period to allow the legislature to act. In short, this was an extraordinary situation requiring an extraordinary solution: paradoxically, allowing the infringement of the rule of law for a period was necessary to respect the rule of law.

Manitoba Language Rights was not, of course, a Charter case. Subsequently, the Supreme Court of Canada has imposed 90 declarations of invalidity, of which it has suspended 23, but none since 2015, although the court has made 13 declarations of invalidity since then (G, Karakatsanis J. for the majority, paras. 118 and 133). Justices Côté and Brown see the statistics somewhat differently:

out of the 44 times this Court has declared a law invalid for unconstitutionality since Schachter, it has suspended that declaration 23 times (that is, 52 per cent of the time). And those numbers have been trending upwards: between 2003 and 2015, that number rose to 74 per cent of declarations…. (G, Côté and Brown JJ., para. 232, citations omitted)

Acknowledging that Schachter established the principles for determining the appropriate remedy in Charter cases, Karakatsanis J. agrees with commentators who argue that it is time to revisit Schachter to develop “a more principled, coherent, and transparent” approach to determining the appropriate remedy for a Charter violation.

[Sidenote: Schachter is an illuminating example of how significant constitutional law can be developed on a knife edge. Although the parties had agreed that the distinction between natural and adoptive parents in the impugned law was unconstitutional and therefore did not provide the Court with the kind of information it usually required (attracting admonition from then Chief Justice Lamer), the Court nevertheless forged ahead to provide the remedy. In G, in a shallow echo of Schachter, Rowe J. refers to the fact that the parties did not “focus[] their submissions on the suspension” (as well, the Court of Appeal did not stay the suspension) and he therefore finds no need to decide whether it was appropriate to suspend the declaration (G, Rowe J. para. 214). The other judges nevertheless went ahead to consider the suspension issue.]

Schachter concerned the benefits for which new parents were entitled under the then Unemployment Insurance Act, 1971. Mothers giving birth were eligible for maternity benefits of 15 weeks, while adoptive parents were eligible for 15 weeks following placement of their child with them, to be shared between them. Schachter had planned to take time off to be with his newborn child once the child’s mother returned to work, but he was ineligible for “maternity” benefits and so applied under the provision relating to adoptive parents. He was denied because he was “not available for work”.

Given the parties’ agreement, the Court declared the provision invalid and suspended the declaration, a course reinforced by the fact that Parliament had already amended the legislation in a manner different from the result the Court’s reading in would have imposed. In the course of determining the appropriate remedy in Schachter specifically, the Court reviewed the issue of appropriate remedy generally, explaining the principles to be applied. It concluded that there were three circumstances under which a suspension of a declaration of invalidity would be warranted: conflict with other constitutional principles, public safety and underinclusive provisions.

Nearly thirty years later, the Supreme Court reassessed Schachter in G, with all judges except Rowe J. expressing dissatisfaction with it or its ballooning aftereffects. For the majority, this means the Schachter principles need clarifying, while for Côté and Brown JJ., it is time to step away from Schachter and return to Manitoba Language Rights and refine the circumstances under which a suspension would be appropriate.

Justice Karakatsanis concludes the courts should be guided by four rather general and anodyne principles:

A. Charter rights should be safeguarded through effective remedies.

B. The public has an interest in the constitutional compliance of legislation.

C. The public is entitled to the benefit of legislation.

D. Courts and legislatures play different institutional roles. (G, Karakatsanis J. for the majority, para. 158)

Thus Karakatsanis J. finds that while Schachter

has provided helpful guidance on how to craft a responsive and effective remedy for unconstitutional laws for nearly three decades [,] in some respects, this Court’s remedial jurisprudence has moved beyond Schachter. By employing and building on Schachter’s guidance in determining the form and breadth of declarations of invalidity, suspending the effect of those declarations, and exempting individuals from suspensions, this Court’s remedial practice has come to coalesce around a group of core remedial principles. Recognizing those remedial principles and explicitly identifying approaches that strike the right balance between them will encourage greater consistency and transparency in remedial decision making. (G, Karakatsanis J. for the majority, para. 82)

Justice Karakatsanis’s central premise is that

suspensions of declarations of invalidity should be rare, granted only when an identifiable public interest, grounded in the Constitution, is endangered by an immediate declaration to such an extent that it outweighs the harmful impacts of delaying the declaration’s effect. And when declarations are suspended, granting individual exemptions pursuant to s. 24(1) of the Charter will often balance the interests of the litigant, the broader public, and the legislature in a manner that is “appropriate and just”. (G, Karakatsanis J. for the majority, para. 83)

Courts should not exercise complete discretion or apply a “rule-based” approach that can become mechanical in application. (G, Karakatsanis J. for the majority, para. 91)

Justice Karakatsanis explains the significance of the first step in determining a remedy: it is necessary to “defin[e] the extent of the constitutional defect by reference to the substantive violation of the Charter” both to protect the rights of those affected by the law and in ensuring that government acts constitutionally. Remedies are not limited to the claimant and may apply even if the claimant is not directly affect because “[n]o one should be subjected to an unconstitutional law”. (G, Karakatsanis J. for the majority, para. 109)

Having determined the extent of the unconstitutionality, the court must decide “the form that a declaration should take”. In addition to declarations of invalidity, the court might read down unconstitutional provisions, read in words that would make the provision constitutional by expanding its application or sever unconstitutional elements (“where the offending portion is set out explicitly in the words of the legislation”). The risk with these remedies is that the court might be substituting its view for that of the legislature in a way that is inconsistent with what the legislature would in fact do and that is consistent with the purpose of the legislation.

She maintains that “[a] balancing approach permits courts to engage with the underlying principles and ensure that a delayed declaration is not ordered unless there are compelling reasons to do so. The appropriate balance will result in suspensions only in rare circumstances”. (G, Karakatsanis J. for the majority, para. 132) Furthermore, “the government bears the onus of demonstrating that the importance of another compelling interest grounded in the Constitution outweighs the continued breach of constitutional rights. In each case, the specific interest, and the manner in which an immediate declaration would endanger that interest, must be identified and, where necessary, supported by evidence.” (G, Karakatsanis J. for the majority, para. 133) The government also bears the onus of establishing the appropriate length of the suspension.

Justice Karakatsanis considers individual constitutional exemptions to be appropriate if doing so safeguards constitutional rights; it is legitimate to grant the applicant an exemption when others in similar circumstances will continue to be subject to the law because the applicant brought the challenge. Indeed, “an individual remedy for the claimant will often be appropriate and just. The importance of safeguarding constitutional rights weighs heavily in favour of an individual remedy” and “there must be a compelling reason to deny the claimant an immediately effective remedy”. However, if granting an exemption would endanger public safety or if it would require extensive use of judicial resources to assess a large group of similarly-situated claimants, the court should not grant the exemption. (G, Karakatsanis J. for the majority, paras. 147, 149 and 150-151)

In G, “[t]he legislature would no doubt have enacted Christopher’s Law to apply to those convicted of a sexual offence even if it could not include those found NCRMD and granted a discharge.” Thus it is appropriate to read the law down to exclude those found NCRMD, “vindicat[ing] rights without interfering with aspects of the statute’s operation unaffected by the finding of unconstitutionality”. (G, Karakatsanis J. for the majority, paras. 164 and 166) In this case, given the need for public safety and legislative discretion in responding to the unconstitutional provision, the majority suspends the declaration. (It must be said that this portion of the judgement is somewhat confusing and if anything, makes it evident that if the reason for suspension is not strong, the declaration should not be suspended.) Justice Karakatsanis grants G an exemption, given his specific circumstances and positive conduct. (G, Karakatsanis J. for the majority, para. 182)

While more stringent in their criticism of how Schachter has been applied (or abused), and without saying so, Justices Côté and Brown share Karakatsanis J.’s view that the court has tended to suspend declarations of invalidity more recently, sometimes without clearly explaining why, but their endpoint is very different. They distinguish their position by describing Karakatsanis J.’s as “represent[ing] an unbridled expansion of judicial discretion”. (G, Côté and Brown JJ., para. 294) Rather than trying to bring some order to how Schachter has been applied, they effectively abandon it, going back to Manitoba Language Rights. They maintain that only a threat to the rule of law justifies a suspension. Since the Manitoba Language Rights decision,

this Court has lost its way. The Court now suspends declarations of invalidity almost as a matter of course, often with no justification or attention to the rule of law. In our view, most of the cases in which suspended declarations have issued since Manitoba Reference do not come close to reaching the high threshold it decreed. Rather, suspended declarations have become this Court’s “remedial instrument of choice”, applied “casually” and as a matter of “routine” or “preference” while affording only “lip service . . . to the dangers of allowing continued violations of Charter rights and freedoms”. (G, Côté and Brown JJ., para. 229)

Rather bizzarely, Côté and Brown JJ. equate (for analytical purposes) suspended declarations to the availability of section 33 of the Charter, the override clause, suggesting, they say, that it indicates that legislatures are better positioned to decide when a suspension is appropriate. Justice Karakatsanis ably explains at paragraph 137 why this comparison is “unsustainable”:

Section 33 permits Parliament or a provincial legislature to temporarily exempt an Act from the application of rights and freedoms guaranteed by ss. 2 and 7 to 15 of our Charter, even for purely political reasons … When a court determines that a law violates the Charter in a manner that cannot be justified in a free and democratic society under s. 1, the court must grant the appropriate remedy. This includes, in some rare cases, delaying the effects of a declaration of invalidity based on a compelling public interest. Court‑ordered suspension leaves Parliament and the legislatures free to respond to the declaration of invalidity, including by using s. 33 …. The court cannot shirk its responsibility to remedy constitutional violations simply because s. 33 permits Parliament or a legislature to exceptionally override certain Charter rights and freedoms. (G, Karakatsanis J. for the majority, para. 137)

Although Côté and Brown JJ. provide a considerable list of reasons why a suspended declaration is inappropriate except when it is necessary to maintain the rule of law and, by extension, when failure to maintain the rule of law would result in a danger to public safety, they are willing to suspend the declaration in this case because otherwise “the Christopher’s Law registry would not apply to all persons found NCRMD who have been granted absolute discharges by the Ontario Review Board”, since they accept the evidence that persons NCRMD are more likely to reoffend. (G, Côté and Brown JJ., para. 268) There is no reason for a constitutional exemption, however, since, among other reasons, the legislature is in a better position to determine how to address persons found NCRMD.

Justice Rowe, while acknowledging that courts have not always respected Schachter’s boundaries, sees no reason to depart from it, for the categories it identifies for suspending a declaration (public safety, conflict with other constitutional principles and underinclusive provisions) are workable and not difficult to apply. Thus “em>Schachter did not prove unworkable in practice; the problem was that it was too often honoured in the breach. It had been tested, it had worked well, and it can continue to do so.” (G, Rowe J., paras. 201-202) Describing Karakatsanis J.’s approach as “discretionary”, Rowe J. expresses his fear that it “will lead to a continuation of current trends, in which declarations of invalidity are suspended in a way that varies with the length of the Chancellor’s foot”, while Schachter “restrains discretion”. (G, Rowe J., para. 205)

Justice Rowe sees no approach to section 52 that will allow judicial discretion. Rather, legislative provisions are contrary to the constitution or they are not: “Courts have no inherent authority to make legal that which is not. In this I differ fundamentally with my colleague.” (G, Rowe J., para. 206) Contrary to the majority’s view, there is no inherent authority in the court that can override the express provisions of the constitution.

Justice Rowe’s criticism of Karakatsanis J.’s principles (as compared with Schachter‘s categories) suggest he would limit the circumstances under which a declaration of invalidity should be suspended (or other remedies employed?), yet he disagrees with Côté and Brown JJ.’s approach because they limit the circumstance under which the court can suspend a declaration of invalidity to the rule of law. For Rowe J., “underinclusive benefits call for a court to order a suspension so as to give proper place to the legislature in framing a remedy.” (G, para. 209)

Accordingly, the risk to public safety does not justify a suspension unless it rises to imperil the rule of law and therefore would not have been appropriate here. In any event, both the suspension and the constitutional exemption are moot, the latter because the legislation is of no force and effect. However, Rowe J. aligns with Côté and Brown JJ.’s approach to constitutional exemptions more generally.

The division among the judges in Fraser indicates a fundamental difference on how to read section 15(1) of the Charter. Fraser indicates why that matters, while G tells us that in the most obvious cases it might not always have an effect. This difference in section 15(1) analysis reflects a difference in how the justices view their role in the constitutional regime. G, too, reveals important distinctions in how the judges think about the scope of their constitutional role. To some extent, this plays out in the case itself, but it has to be said that as we leave G, we are left, on the one hand, with little real guidance beyond Schachter despite the majority’s efforts to clarify it and, on the other, with strong aversion to an expansive role for the courts reflected in the opinions of Côté and Brown JJ. and Rowe J.