Democracy and the Notwithstanding Clause

Abstract

This article focuses on the relationship between democracy and the notwithstanding clause in s.33 of the Canadian Charter of Rights and Freedoms. A number of scholars argue that s.33 is inherently ‘democratic’, as it is an assertion of legislative supremacy. The most influential such theory is Jeremy Waldron’s. This article offers a democracy-based critique of Waldron’s democracy-based account of the notwithstanding clause. The argument that the notwithstanding clause is necessarily ‘democratic’ ignores the constitution of the legislature through elections and the risk of self-dealing by the legislative branch, adopts an idealistic view of legislatures at odds with the reality of executive dominance and party discipline, and over-relies on the assumption that the electorate will ensure retrospective accountability for misuse of s.33. Contrary to Waldron and those who have adopted his arguments in Canada in the context of the Charter, the article argues we can be democrats and have faith in the capacities of legislators and voters while still maintaining skepticism about the uses to which the notwithstanding clause may be put. In short, s.33 is not inherently democratic. The political morality of each use of the notwithstanding clause—including whether it helps or harms democracy—must be assessed on a case-by-case basis.

I. Introduction

While the ‘notwithstanding clause’ in s.33 of the Canadian Charter of Rights and Freedoms ¹ “has emerged as [the Charter’s] most controversial provision,” the death of the clause has been proclaimed many times since 1982. ² Outside of Quebec, ³ s.33 had been invoked sparingly until recently. ⁴ The degree of political restraint reached the point where scholars began to consider whether s.33 had fallen into “desuetude,” ⁵ if there was a constitutional convention against its use outside of Quebec, ⁶ or if it was “simply irrelevant.” ⁷ The era of political restraint appears to be over. ⁸ Ontario invoked s.33 for the first time in 2021 ⁹ and openly considered using it in another case, ¹⁰ both times in relation to political rights. Quebec’s recent use of s.33 in legislation on state secularism ¹¹ and language rights ¹² has been particularly controversial.

This article focuses on the relationship between s.33 and democracy in general and the use of the notwithstanding clause in relation to political rights in particular. A number of scholars have advocated for more frequent use of the notwithstanding clause on explicitly democratic grounds. ¹³ While the debates about the notwithstanding clause have often centered around the separation of powers, dialogue theory, and so on, proponents of the use of s.33 have rooted their arguments in the democratic legitimacy of the legislative branch, namely Parliament and the provincial legislatures. The most influential theory of the legitimacy of the notwithstanding clause is Jeremy Waldron’s democracy-based account. ¹⁴ Waldron’s reasoning has been directly adopted by some who argue for the democratic legitimacy of s.33. ¹⁵

This article offers a democracy-based critique of Waldron’s democracy-based account of the notwithstanding clause. ¹⁶ In brief, Waldron’s argument as applied to s.33 goes as follows. ¹⁷ In a society with inevitable disagreement about moral principles, the right of political participation ensures each citizen has a say in matters affecting their interests. Most importantly, it demonstrates the respect owed to each individual as a political equal capable of deliberating about politics, including the substance of rights. Political participation in a representative democracy culminates in the institution of the legislature, which has superior democratic legitimacy in comparison with the unelected judiciary. Democracy requires that the legislature be the venue for resolving conflicts about rights, including political rights. In this argument, legislative invocation of s.33 possesses an unimpeachable democratic pedigree. Allowing courts to decide moral matters or to resolve disputes about constitutional rights is a repudiation of political equality and is, therefore, anti-democratic. In the most robust version of this argument, concern about potential misuse of s.33 reflects undue cynicism as to the intentions of legislators and an elitist tendency to look down on voters by doubting their capacity to deliberate. Skepticism of s.33 has therefore sometimes been characterized not as trepidation about the potential misuse of state power, but as revealing reticence about democracy itself. ¹⁸

This article argues that there are contradictions at the heart of the Waldron-ian account of democracy and s.33, as the notwithstanding clause itself can be used to subvert democracy. Waldron’s argument is explicitly conditional on healthy democratic institutions already being in place and the right of political participation, which he calls the “right of rights,” being well-protected. ¹⁹ Such preconditions cannot be assumed, even in a country with a long democratic tradition such as Canada. The actual as opposed to the ideal operation of the main institutions of constitutional democracy, especially the legislature, must be assessed. The scope of s.33 also leaves significant room for the “right of rights” to be unduly restricted. Even though s.33 is inapplicable to the right to vote (s.3) and some structural democratic rights (ss.4-5), it applies to freedom of political expression and association (s.2) as well as equality rights (s.15), which are all directly relevant to elections. If Waldron is correct that invocations of s.33 are legitimate because of the democratic nature of a representative legislature, then uses of the notwithstanding clause that harm electoral accountability or political participation should be seen as illegitimate even on his own account. They are illegitimate for being inconsistent with the very reason that he argues we should respect the legislature in the first place. A truly democratic account of the notwithstanding clause must accept, in other words, limits on its use where the right to political participation, electoral accountability, or democracy writ large is in peril. The democratic legitimacy of any particular use of s.33 must be assessed in this light.

This article proceeds as follows. Section II sets out Waldron’s defence of s.33 on democratic grounds, including the centrality of political participation to his theory. Section III articulates the democracy-based critique of s.33. I will argue that Waldron’s account: 1) ignores the constitution of the legislature itself through elections; 2) assumes an ideal legislature with robust checks on political behaviour; and 3) over-estimates the capacity of elections to discipline misuse of s.33. As a result, the democracy-based defence of s.33 as articulated to date has failed to satisfactorily address the potential risks to democracy from misuse of the notwithstanding clause. That account is deficient for not recognizing limits on the legitimate uses of s.33. Section IV argues that the potential for s.33 to be used so as to harm political participation or electoral accountability remains a serious concern. In short, while the text of s.33 excludes ‘Democratic Rights’ in ss.3-5 of the Charter from s.33, many of the freedoms that make elections truly free and fair are still subject to the clause. ²⁰ Section V concludes by considering the implications of these arguments.

II. The Democracy-Based Defence of s.33

a) The Waldron-ian Account

The main argument by proponents of more frequent invocation of s.33 by legislative bodies is based in a particular account of democracy. Whether used in response to a judicial decision striking down a statute as invalid on Charter grounds or pre-emptively in anticipation of such a result, ²¹ s.33 allows legislatures to have the final say on the rights and freedoms to which it applies. ²² As Jeffrey Goldsworthy writes, “the most powerful and popular argument against the judicial enforcement of constitutional rights maintains that it is undemocratic for unelected judges to invalidate laws enacted by a democratically elected legislature.” ²³ While the democratic argument for constitutional interpretation by legislatures has several variations, the argument’s most thorough articulation comes from Jeremy Waldron. Waldron has also shaped the Canadian debate. ²⁴ It is therefore worthwhile setting out in some detail Waldron’s claims regarding political participation, judicial review, and the legislature, as well as on s.33.

Waldron starts from the premise that there will be genuine differences in society that must be respected as to how rights should be interpreted. ²⁵ The foundational right in a democracy is political participation, the “right of rights.” ²⁶ Waldron rejects instrumental accounts of the rights to vote and participate in politics. ²⁷ We should instead respect a right of political participation, Waldron argues, because of the “denigration that is involved when one person’s views are treated as of less account than the views of others on a matter that affects him as well as the others.” ²⁸ The “peculiar insult” in denying a right of political participation to a person is due to the impact on the person’s “own rights and interests,” which are affected without their input. ²⁹ Denying political rights means rejecting the excluded individual’s “capacity to decide responsibly” or to deliberate on the matter at issue. ³⁰ The right of political participation stems from the “capacity of ordinary people for intelligent and conscientious moral deliberation.” ³¹ If we are to take seriously the equal moral worth of each individual and their capacities, we must respect their right to participate.

Given the inevitability of conflicting interpretations of rights and clashes between rights, we must collectively decide which institution is the preferred venue for resolving those conflicts. For Waldron, the answer is the legislature chosen by the people exercising their right to participate on equal terms and populated by legislators whose capacity to deliberate should also be respected. Waldron’s “central objection to judicial review is that it denies (or at least curtails) citizens’ ‘right to democratic participation’.” ³² Legislatures are the preferred institution in which to make decisions about “what rights we have.” ³³ It is elitist and anti-democratic to have a “judicial aristocracy” ³⁴ or “scholarly or judicial elite” ³⁵ substitute their view as to how to resolve disputes about rights for that of the electorate channeled through the legislature. Judicial review is “politically illegitimate,” as “privileging majority voting among a small number of unelected and unaccountable judges … disenfranchises ordinary citizens and brushes aside cherished principles of representation and political equality.” ³⁶ His method focuses on procedures ³⁷ for resolving disputes about rights, rather than an instrumental assessment of which institution is better able to provide specific outcomes. ³⁸ Waldron’s claim that legislatures are generally superior to courts for resolving complex moral disputes is therefore based in the right of political participation, the equal moral worth of the individuals who elect the legislature, and the capacity of legislators to deliberate, rather than a weighing of which institution has been historically or is likely to be in the future the best rights protector.

In writing explicitly on s.33, Waldron applies his existing theory of the democratic legitimacy of the legislature based in the right of political participation. Section 33 is taken to be the ultimate articulation of parliamentary supremacy in the Canadian context, though a flawed one. In his view, s.33 does not go far enough in ensuring parliamentary supremacy over the courts. That is partly why Waldron classifies Canada as a system of strong rather than weak form judicial review, despite the notwithstanding clause. ³⁹ Waldron wishes s.33 were used more often, but he sees it as too limited to ultimately foster a fully empowered democratic majority. Waldron therefore critiques the text of s.33 for being worded so as to imply that the legislature is ‘overriding’ the Charter. ⁴⁰ The implication flowing from the text is that the legislature is overriding a particular right or freedom rather than offering its own interpretation that happens to genuinely differ from that of the courts. He also views the time-based limits on its use imposed by the 5-year sunset clause as overly restrictive. ⁴¹ Whatever his critiques of the particular ways in which the text of the Charter limits the reach of the notwithstanding clause, Waldron’s theory lends itself to a robust defence of s.33 on democratic grounds.

One possible response to Waldron’s argument that judicial review is democratically illegitimate is that it was political majorities in Canada who freely chose to entrench a bill of rights. ⁴² If political majorities have democratic legitimacy, then the decision in 1982 to entrench the Charter should be respected as should the subsequent decisions not to invoke s.33. ⁴³ Waldron flatly rejects this line of argument. ⁴⁴ He claims that if a legislature turns itself by legal means into a dictatorship, that doesn’t justify the choice on democratic grounds. Democracy has simply “been extinguished democratically,” as Jeffrey Goldsworthy puts it in summarizing Waldron’s view. ⁴⁵ A democracy choosing an entrenched bill of rights in Waldron’s account means nothing less than “voting democracy out of existence, at least so far as a wide range of issues of political principle is concerned.” ⁴⁶

United by concern for how best to preserve democracy, scholars have taken up versions of Waldron’s core arguments against judicial review in the context of the notwithstanding clause. ⁴⁷ Now-Justice Grant Huscroft adopts Waldron’s reasoning in his academic analysis of the Charter and the notwithstanding clause. ⁴⁸ Huscroft seeks to critique dialogue theory and, therefore, focuses to a large extent on the separation of powers. ⁴⁹ He writes for example that,

I suspect that many proponents of dialogue theory would prefer that important decisions be made by the Supreme Court of Canada rather than Canadian legislatures. They cannot bring themselves to admit this essentially elitist position, however, because democracy has a greater hold on the public imagination than judicial review. For all of its imperfections, democracy affords a voice to the people who are governed, and for their part, the people rightly expect that they should be involved in making the decisions that concern them. ⁵⁰

Huscroft picks up on Waldron’s claim that allowing judicial review on rights grounds undermines the individual’s right to political participation, understood mainly as the right to have a say on matters that affect them.

Dwight Newman also argues for more frequent use of s.33 primarily on what appear to be democratic grounds. He adopts a version of Waldron’s argument, though one that departs from it by setting out some general categories of illegitimate legislative action. Newman critiques extreme skepticism of legislatures’ capacity to deliberate and of s.33 as contrary to “the very idea of democratic governance,” which rests on faith in elected representatives. ⁵¹ Like Waldron, he emphasizes political constraints on s.33: “Parliamentarians and legislatures contemplating the use of the notwithstanding clause must act in a politically responsible manner.” ⁵² He goes beyond Waldron by setting out some specific limits on s.33. He writes that the notwithstanding clause’s “routine” use is objectionable and that it “would not be appropriate to have legislatures excessively involved in second-guessing judicial decisions on an everyday basis.” ⁵³

Jeffrey Goldsworthy’s comparative analysis of parliamentary sovereignty has Canada as a prominent case study, including detailed consideration of s.33. ⁵⁴ He argues that s. 33 is a sign of respect for the wisdom of the electorate, which selects representatives, rather than merely an institutional division of labour granting power over constitutional interpretation to the legislature in some circumstances. To doubt the wisdom of the electorate is anti-democratic in his argument, as on Waldron’s. “[All democrats] should have considerable faith in the inherent intelligence, knowledge and virtue of the electorate.” ⁵⁵ Goldsworthy argues that s.33 offers the possibility of more democratic constitutional politics than one dominated by courts: “In principle, an override clause such as s.33 should help legislators resist the democratic debilitation that might otherwise attend the legalization of rights.” ⁵⁶ It is courts that threaten democracy on his account, not the majority acting contrary to minority interests. Goldsworthy’s account echoes Waldron’s in tying the legitimacy of the legislature to the respect owed to the electorate and in viewing s.33 as the ultimate assertion of the authority of the democratically elected legislature in Canada.

III. A Democracy-Based Critique of the Notwithstanding Clause

a) Introduction

The democracy-based accounts of s.33 set out above generally do not address the threat to democracy from the misuse of the notwithstanding clause in ways that substantially undermine representative democracy. This section lays out a democracy-based critique of the democratic defence of the notwithstanding clause. It makes three main critiques.

First, Waldron’s account of democracy and s.33 ignores the constitution of the legislature through elections. Some laws passed by the legislature using the generally-accepted procedures would undermine the representativeness or democratic legitimacy of the legislature itself. Waldron’s account in particular is explicitly based on several assumptions, which set aside this possibility of legislative misbehaviour. For his theory to be correct, Waldron requires us to assume away the risk that self-dealing legislatures will at times craft election laws that hinder fair terms of democratic competition or restrict the right of rights. This willing suspension of disbelief wishes away some of the most pressing problems for contemporary democracy, including the prospect of a legislature that is composed of an appreciable number of members who are hostile to the institutions and rights that make democracy possible.

Second, the legislature in the democratic account of s.33 is envisioned in an idealized fashion, rather than assessed on its actual performance. The critique that Waldron views the legislature in overly optimistic terms has long been a standard one. ⁵⁷ What is worthy of further elaboration in the context of s.33 is the degree to which Waldron’s account rejects any evaluation of the actual performance of legislatures, courts, and executives in securing or undermining rights. The many critics of judicial review post-Charter have catalogued in great detail the failings of courts and how they actually operate. The defenders of s.33, however, do not engage in a similarly skeptical manner with the actual operation of the legislature. They are idealists about legislatures, but realists about courts. They are following Waldron in that respect, who mounts an abstract defence of legislatures as an institution. I argue instead that it is unpersuasive to adopt a one-sided account of institutions. In the context of s.33 we need to assess the likelihood that the legislature will act irresponsibly, even if it acts responsibly most of the time on most issues. The gravity of s.33’s potential misuse for Canadian democracy requires no less. ⁵⁸ Even though s.33 does not apply to the democratic rights in ss.3-5 of the Charter, there is still much room for it to have an impact on political participation through s. 2 and s.15.

Third, the defences of s.33 as democratic require heavy reliance on electoral accountability as disciplining governments and legislatures. The text of s.33(3) imposes a 5-year sunset clause on invocations of the notwithstanding clause, which facilitates periodic electoral accountability since there must be a general election at least every five years. ⁵⁹ If the courts cannot check legislatures because s.33 has been invoked, then the electorate can, as the theory goes. The implication of that claim is that if misuses of s.33 can be checked at the ballot box, then there is less reason to worry about the notwithstanding clause being used to undermine democracy. For a variety of reasons, however, elections are imperfect accountability mechanisms if the goal is retrospective accountability. The realistic prospect of the sunset clause facilitating electoral accountability should not be over-stated.

If the critiques of Waldron’s account have salience, then uses of s.33 which harm electoral democracy are illegitimate, even on Waldron’s own reasoning. They contradict the very reason that Waldron’s account tells us that s.33 exists. Waldron’s theory is unable to justify limits on political rights because of the foundational nature of political participation to his account. The legislature has legitimacy by virtue of being chosen by free and equal citizens exercising their right to political participation. The right of rights has special primacy in Waldron’s account. On Waldron’s logic, if s. 33 in the Canadian constitutional order is the ultimate assertion of legislative sovereignty, then it is legitimate because of its democratic pedigree. If the legislature uses the notwithstanding clause in a manner that harms the right of political participation or electoral competition, however, then it forfeits that legitimacy. Using s.33 to harm the right of rights means denying equal concern and respect to individuals who have a stake, because their rights or interests are affected, and who have a capacity to deliberate. The reason for respecting the legislature as the institution in which to resolve disputes about rights, on Waldron’s own argument, dissolves if the legislature opts to thwart the right of rights or was elected under conditions that were less than free and fair. Under such conditions, the legislature cannot be assumed to be the institution which best reflects respect for rights holders.

Most importantly, his theory fails to provide resources that would allow us to distinguish democracy-enhancing from democracy-harming actions by the legislature. Waldron assigns political legitimacy to institutions, rather than focusing on outcomes for rights holders, and then deliberately ignores the actual functioning of those bodies. Waldron’s idealized portrait of legislatures therefore incorporates no concrete limits on how they should use their authority. In his verve to mount as robust a defence of legislatures as possible, he fails to give due credence to the basic reality that any state power can be misused. One of the central challenges for constitutional democracy in Canada in a world where the notwithstanding clause is regularly used is how to assess the political legitimacy of those exercises of legislative power. Waldron’s theory does not provide guidance.

b) The Constitution of the Legislature Through Elections

One of the defining and enduring challenges of contemporary democracy is the problem of self-dealing by legislatures. Representatives are elected by the laws established by the legislature to structure elections. Elected representatives compete in the game, but also shape its rules. There is an ongoing risk, therefore, that legislatures will design the rules of the game to favour themselves in some way, rather than acting public-mindedly in legislating the operation of elections. ⁶⁰ Whether laws are crafted to facilitate a particular partisan outcome or to protect incumbents, there is plentiful evidence of self-dealing on election laws in countries with long democratic track records, including parliamentary democracies. ⁶¹

If there is a risk of behaviour by the legislature that targets political rights, insulates its members from the full force of electoral accountability, and so on, then the presence of s.33 exacerbates the challenge. Contemporary democracies have a variety of accountability mechanisms applicable to those wielding public power. Elections, political checks within the legislature and executive including conventions of responsible government, review by the courts, dissent expressed within political parties, independent institutions such as the election commission, the media, and so on all act to discipline political behaviour. Elections are not the only accountability mechanism, but they are an indispensable and defining one for democracy to exist. Section 33 provides the opportunity for the legislature to undermine electoral accountability by stymying judicial invalidation of laws that would otherwise be found to violate democratic rights.

Waldron assumes away this problem by relying on four explicit assumptions in his argument on the merits of legislatures as the preferred institution for resolving disputes about rights. Waldron’s argument assumes the presence of: 1) “democratic institutions in reasonably good working order, including a representative legislature elected on the basis of universal adult suffrage”; 2) judicial institutions “in reasonably good order”; 3) a societal commitment to individual and minority rights; and 4) “persisting, substantial, and good faith disagreement about rights … among the members of the society who are committed to the idea of rights.” ⁶²

Adopting these assumptions is tantamount to assuming away the strongest objections to his democracy-based critique of judicial review. These assumptions allow Waldron to say that even if courts and legislatures are both operating in “good order,” legislatures remain the preferred and most legitimate institution in which to resolve disputes about rights. They also mean, however, that he does not address breaches of political morality by the legislature, including legislative action that harms the right of rights. ⁶³ In the Canadian context, the concrete risk is that legislative supremacy facilitated by regular use of s.33 and unchecked by the courts or other institutions could lead to the diminution of minority rights of political participation. ⁶⁴

Waldron does provide further details regarding what he means by the assumption of healthy “democratic institutions.” ⁶⁵ He says for the assumption to hold, elections must be “fair and regular.” ⁶⁶ He assumes the presence of bicameralism, deliberation within the legislative bodies, internal procedures for passing laws, the presence of parties, and so on. He tries to set aside any critique of his account by stating: “None of this is meant to be controversial; it picks out the way in which democratic legislatures usually operate.… I am assuming that the democratic institutions are in reasonably good order. They may not be perfect.” ⁶⁷ He then goes on to stipulate a culture that respects “political equality” as part of his argument. ⁶⁸

None of these details address the prospect of unhealthy democratic institutions or the particular problem of self-dealing. The closest he gets to acknowledging this critique is to write that, “It is sometimes said that elective institutions are incapable of reforming themselves because legislators have an entrenched interest in the status quo.” ⁶⁹ He acknowledges that this concern may be a relevant one, but confines its salience to the United States and claims, “It is patently false elsewhere.” ⁷⁰ The only example he provides to substantiate this general claim about democracy globally is one particular example from New Zealand, namely the transition from a first past the post to a mixed-member proportional system (MMP) in the 1990s. ⁷¹ There is plentiful evidence in the political science literature, however, that choice of electoral system is one of the areas of legislative or government authority where self-dealing is most clearly evident. ⁷² The evidence isn’t that legislatures are always bad faith actors, but that self-dealing by legislatures is a pervasive and ongoing problem that Waldron unduly downplays.

The right of political participation is so integral to Waldron’s theory that he argues it should not be limited to further other rights or principles. ⁷³ Section 33 allows legislatures a nearly unfettered hand to restrict political participation, however, with the exception of the right to vote. I agree with Waldron that the right of rights is indeed special and requires unique protection, given that it is foundational to self-government. Waldron’s insistence on a pure theory of legislative supremacy, however, amounts to unilateral disarmament in the ongoing struggle to preserve democratic governance. Rather than seeking to check potentially serious consequences for democracy of its misuse, Waldron’s theory simply asks us to trust the legislature. A defence of s.33 on the basis that legislatures always have democratic legitimacy in all of their actions ignores the serious consequences for democracy if the notwithstanding clause is misused to limit political participation and free and fair elections. It also stands in contradiction with Waldron’s own adoption of the right of rights as the centrepiece of his theory of institutional legitimacy.

c) The Non-Ideal Legislature

The advent of the Charter forced Canada to wrestle with the presence of a written bill of rights enforceable by the courts alongside the tradition of parliamentary sovereignty. ⁷⁴ With newfound authority granted by the framers and embraced by the judiciary, the courts have deservedly come under intense scrutiny in the 40 years of living with the Charter. The advocates for more frequent use of the notwithstanding clause have been unsparing in their criticism of the actual operation of courts, as opposed to their theoretical role in rights protection. Legislatures deserve equal scrutiny, especially if they are going to be more forceful in exercising the powers granted to them in s. 33. We should not be realists about courts but idealists about legislatures. ⁷⁵

While writing in the political constitutionalist tradition, ⁷⁶ Waldron’s lack of skepticism with regard to the legislature is at odds with one of the leading lights of that line of thought, J.A.G. Griffith. ⁷⁷ Griffith in his famous “The Political Constitution” argues for healthy skepticism toward all institutions wielding public power, the legislature included. ⁷⁸ Tom Poole writes that,

‘The Political Constitution’ is no paean for democracy, then, but is better read as a diatribe aimed at instilling a rigorous scepticism into constitutional thinking. We should—indeed, if we value our freedoms, must—be sceptical in the face of all claims and justifications for power, whether the claim comes from a judge [or a] politician. ⁷⁹

The Waldron-ian lack of skepticism toward the legislature is reflected in the writing of Canadian scholars who apply categories to understand how the notwithstanding clause has been used to date. Janet Hiebert sets out four categories for how s.33 has been invoked. ⁸⁰ The first is political protest in the early years of the Charter, particularly Quebec’s “omnibus and retroactive” use of s.33 from 1982-1985. ⁸¹ Quebec purported to apply the notwithstanding clause to all legislation, which led to the decision of the Supreme Court of Canada in Ford v Quebec in 1988 setting procedural limits on s.33’s invocation. ⁸² The second and third are forms of risk aversion in the face of uncertainty about how the courts would interpret the Charter. ⁸³ Such risk aversion can lead legislatures to pre-emptively use s.33 even prior to the invalidation by the courts of a statute for non-compliance with the Charter. Hiebert provides examples mainly from the early years of the Charter as evidence of the relevance of these categories. ⁸⁴ The fourth and most important category is political disagreement over judicial interpretation of the Charter. It is here that Parliament or a legislature would offer a different interpretation of a Charter right or freedom than do the courts. The now much-recited argument that it is a misnomer to label s.33 as the ‘override’ provision rests primarily on this category. On this theory, the legislature is merely offering a different interpretation of a provision.

While not explicit on the underlying normative assumptions, Hiebert like Waldron does not address the possibility that the legislature functions less than ideally. Hiebert’s four categories do not include one for misuse of s.33 in regard to political participation rights or otherwise. That legislatures do not always behave in a perfectly deliberative fashion, or that they may engage in self-dealing in setting election law, or that they may discriminate against an unpopular minority are simply not part of the categories. Hiebert does draw a distinction between constitutional interpretations put forward by Parliament in “good faith,” as opposed to “rolling the dice” where the law seems certain to be ruled unconstitutional. ⁸⁵ That distinction, however, goes to the likelihood of a law being invalidated by the courts, not whether the legislature is acting in good faith. Hiebert proposes technical changes to the role of the Minister of Justice and Parliamentary committees in assessing possible invocations of s.33 and constitutional interpretations. ⁸⁶ While such reforms have merit as a general matter, it is hard to see them acting as a robust enough check against a determined government.

A potential response to skepticism of legislatures is that the public-mindedness of responsible legislators protects against misuse of s.33. Waldron argues that we should have faith that “voters and legislators are capable of focusing their deliberations on the general good.” ⁸⁷ Concern for the possibility that s.33 will be misused, however, doesn’t turn on an anti-democratic rejection of the capacity of voters and legislators. As Aileen Kavanagh put it in an early critique of Waldron: “Capacity for moral judgement is not inconsistent with moral fallibility.” ⁸⁸

There are also structural reasons to worry about the actual as opposed to idealized performance of Canadian legislatures, particularly around checking governments and engaging in deliberation about rights. Canadian politics is characterized by high levels of party discipline, executive dominance, and the centralization of power in the Prime Minister or Premier’s office. ⁸⁹ There is plenty of evidence that shows that political checks within legislatures are not as robust as they should be. The House of Commons may also operate very differently than a provincial legislature with no upper house, less media scrutiny, and so on. While legislators of course have agency, to be blunt they are often closely directed by the executive.

The role of the executive is generally under-theorized in Waldron’s considerable oeuvre. For example, he does not engage with judicial review of executive action in administrative law, which is generally seen even by political constitutionalists as legitimate. ⁹⁰ Most importantly in relation to political legitimacy, Waldron fails to account for a strong executive that in contemporary democracies tends to dominate the legislature. Many other leading constitutional theories have not adequately engaged with the reality of executive dominance of the legislature. ⁹¹ For Waldron, it is a particularly serious omission. To the extent that the executive dominates the legislature, then the hierarchy of political legitimacy that he advances with the legislature at the top is less persuasive. If what is formally legislative action is tightly controlled by the executive in practice most of the time on most issues, then the search for political legitimacy is not purely a contrast between the courts and the legislature, but between the courts and the executive as well. If it is not the elected executive but its unelected components that direct the government, Waldron’s claims are even harder to sustain. Put bluntly, the question is not always, “Should nine unelected lawyers in robes or the elected legislature decide?” It can also be, “Should unelected political staffers in the Prime Minister’s Office decide?” The answer to such a question is messier than the artificial zero-sum competition over political legitimacy set up by Waldron between courts and legislatures.

One of the early defenders of s.33, Peter Russell, qualified his support for the provision on an evaluation of how legislatures actually operate. ⁹² He argued that while over-reliance on judicial review should be critiqued, support for the notwithstanding clause should be conditional on it being “properly used,” by which he meant “only after a reasoned debate in the legislature.” ⁹³ While Russell was optimistic that Canadian legislatures mostly behaved deliberately, ⁹⁴ more recent studies have shown that there is little evidence of serious deliberation regarding the notwithstanding clause within legislatures when s.33 is invoked. ⁹⁵

Some advocates of s.33 have responded to concerns about the functioning of the legislature by dismissing them as the fault of courts and the Charter. To the extent that legislatures operate in a less than ideal fashion, that is because courts through judicial review have usurped the legislature’s constitutional role, ⁹⁶ therefore leading to ‘democratic debilitation’. ⁹⁷ The argument makes the claim that while the deliberative features of legislatures as interpreters of rights may have diminished in response to judicial supremacy post-Charter, legislative capacity can be built back up in short order. ⁹⁸ Normalizing the use of s.33 would lead to the resuscitation of legislatures.

To say that the failings of legislatures are the fault of courts is unpersuasive. The argument assumes without providing evidence that the atrophying of the legislature’s role in interpreting the Constitution is primarily a result of the Charter itself and the expansion of judicial power. Substantiating the causal claim that the Charter has led to democratic debilitation would require establishing some golden age for Parliament and the provincial legislatures prior to 1982 characterized by deep deliberation about rights compared with the post-1982 reality. No such evidence has been provided. Part of the motivation for an entrenched bill of rights in the first place was the uneven commitment to rights by legislatures and governments prior to 1982. Even without the Charter, structural barriers to meaningful deliberation on rights by the legislature would remain, such as party discipline, executive dominance, and the role of unelected political staffers.

It is important to be clear that I am in no way arguing that only legislatures and executives can behave anti-democratically. Judges sometimes make errors, issue decisions shaped by their personal political preferences, defend the institutional interests of the judiciary, or, even, act with hostility toward democracy. There are prominent instances of courts harming the right of rights. For example, the United States Supreme Court’s current majority is hindering democracy and deliberately tilting the electoral playing field. ⁹⁹ Waldron’s warning that courts often err and do not always protect rights remains powerful, especially in the context of judges acting as partisans. ¹⁰⁰ The possibility that a court might behave anti-democratically is certainly one that must be taken seriously.

A default view that legislatures protect rights and that any decision-making by courts is unavoidably anti-democratic, however, is one-sided. A more even-handed approach than Waldron’s should recognize that democracy is always vulnerable and that there are multiple pressure points. All institutions wielding public power are fallible and should be scrutinized. The scale of the likely risks to democracy in any given system must be carefully evaluated. Part of that assessment is a context-specific one about which branch is more likely to protect rights and which is more likely to hinder their realization.

In some of his more recent work, Waldron has conceded that political majorities may not always be reliably on the side of rights protection for some unpopular minorities, for whom the courts are the best institutional ally. He argues that “criminals and criminal suspects” can be seen as the “beneficiaries of judicial review” in New Zealand, for example. ¹⁰¹ Waldron is here relaxing his famous four assumptions and accepting some of the basic premises of political process theory espoused by John Hart Ely ¹⁰² and encapsulated in the famous Footnote #4 from the decision of the United States Supreme Court in Carolene Products. ¹⁰³ Carolene Products opined that judicial review should protect “discrete and insular minorities” and Ely was particularly concerned with the political rights of such groups. ¹⁰⁴ While expressing reservation about making too much of the point, Waldron concedes that “any general account of the value or otherwise of judicial review ought to include a sober assessment of those cases (whether many or few) where political circumstances make it unlikely that minorities will be able to secure their rights through the ordinary political process.” ¹⁰⁵

If the political process regularly fails certain minorities, then we should certainly be attuned to the possibility that s.33 can and will be used in a way that harms minority rights. If the political process is likely to fail them and the courts are barred from intervening by s.33, then minorities will be in a distinctly precarious position. It is hard to sustain the argument that all uses of s.33 are ‘democratic’ if used to harm the political rights of discrete and insular minorities. Viewed in this light, the potential misuse of s.33 is not a boutique concern or remote possibility, but a defining challenge given the particular make-up of Canadian constitutional democracy.

d) Imperfect Electoral Accountability

The account of the notwithstanding clause as having impeccable democratic credentials also rests on a particular set of assumptions as to how elections operate. Arguments that the notwithstanding clause should be used more regularly tend to emphasize that the public gets to pass judgment at the ballot box on invocations of s.33. If a government and legislature choose to attack the political rights of a minority group and shield it from judicial review by use of s.33, the wisdom of the disapproving electorate remains as the ultimate check.

Waldron and Goldsworthy’s deep faith in the wisdom of the electorate as a collective and in the capacity of individual citizens to deliberate is the foundation of their defence of political majorities. The electorate’s role is to deliberate about rights and public policy and to elect representatives who will exercise their judgment. ¹⁰⁶ Elections are the ultimate accountability mechanism if elected representatives err and, on this theory, legislatures can be punished for misuse of s.33 by voters. ¹⁰⁷ In sum, the account of s.33 as democratic relies on an assumption that elections are mechanisms of retrospective political accountability.

This account finds resonance in the text of s.33. Section 33(3) states that a declaration made under s.33(1) “shall cease to have effect five years after it comes into force.” ¹⁰⁸ Five years is the maximum duration of Parliament or a provincial legislature under s.4(1) of the Charter. Section 33(3) therefore ensures the electorate has the opportunity to pass judgment on the legislature’s invocation of s.33 prior to its expiration. ¹⁰⁹ Section 33(3) has little impact, however, if the notwithstanding clause has been used to diminish the political rights of those likely to oppose the invocation of s.33. The idea of elections as mechanisms of retrospective accountability for misuse of s.33 evaporates if the notwithstanding clause is used in a fashion that harms the right of political participation to a non-trivial extent.

Imagine a legislature that banned a political party in a province. Such actions could plausibly be seen by the courts as being primarily a matter of freedom of association, which is protected in s.2 of the Charter and subject to the notwithstanding clause. ¹¹⁰ The capacity of the electorate to fulfill its constitutional role in holding a government and legislature to account would be hindered if that party had a substantial level of support. If the political rights of the electorate are reduced through invocation of s.33 and the harm is beyond the reach of courts to remedy, it will be of great consequence whether the restriction is significant enough to hinder electoral accountability.

Robert Leckey and Eric Mendelsohn offer a more compelling view on the role of the electorate in relation to s.33. In their assessment, s. 33(3) makes clear that the “legitimacy [of legislatures] in matters concerning the notwithstanding clause derives from democratic support” and that “the use of the notwithstanding clause requires the electorate’s ongoing, or at least episodic, democratic consent.” ¹¹¹ They rightfully encourage us to focus on “the electorate’s ability to play the constitutional role assigned to it by subsection 33(3).” ¹¹² The capacity of the electorate to fulfill its constitutional role may in fact be hindered by uses of s.33 that harm political participation at a non-trivial level.

Elections are also not necessarily straightforward mechanisms of retrospective accountability. The electorate must consider multiple factors in its deliberations. The electorate sits in judgment of the decision to invoke s.33, the justifications on offer for doing so, the quality of deliberation in the legislature on the matter, the impact on rights and freedoms, and the alternate vision of the constitution that the legislature substitutes for the actual or anticipated one of the courts. Even where voters may consider an invocation of s.33 to be abhorrent, however, the prospect of imposing retrospective accountability may reasonably be outweighed by other considerations viewed from the perspective of an individual voter. Voters may favour re-electing the government when considering alternate possible governing parties. Other issues than s.33 may dominate the political campaign and seem the most important basis upon which to cast a ballot. There has been no Canadian political campaign in recent memory in which the decision to use s.33 dominated the discussion among voters. In other words, voters may reason prospectively about the government and policies they want in the future rather than focusing on retrospective accounting for past political misdeeds.

Assuming as we should and as Waldron and Goldsworthy would have us do that voters are well-intentioned and highly capable, the electorate may still opt not to punish misuse of s.33 for valid reasons. While it is certainly possible that voters will cast ballots on the basis of s.33, the assumption that elections provide a robust check should not be over-stated. Being skeptical of elections as mechanisms of retrospective accountability does not reflect doubt as to the capacities of voters. Rather, it reflects a nuanced and realistic understanding of how elections actually operate.

IV. The Text of the Charter and the Problem of Scope

While the preceding portions of the article have largely addressed the relationship between s.33 and democracy generally, it is worth considering in more detail the particular ways that the notwithstanding clause can be used in relation to democratic rights and freedoms. This section addresses what one can call the scope of the notwithstanding clause and the potential impact of its more frequent use on democratic rights and freedoms. Section 33 applies only to sections 2 and 7-15 of the Charter, but not to sections 3-5, ¹¹³ which are grouped under the heading of ‘Democratic Rights’. ¹¹⁴ As a consequence, s.33 cannot be used in relation to the right to vote contained in s.3.

One argument available to those who wish to see the notwithstanding clause invoked more frequently is therefore that any potential harms to democratic rights are insignificant, because the text of the Charter precludes the use of s.33 in relation to them. In my view such an argument is unpersuasive. This section makes two main points.

First, the exclusion of sections 3-5 from the scope of s.33 was a meaningful design choice that reveals underlying normative commitments about the need to protect democracy. Recognition that the notwithstanding clause can be misused in a manner that harms political participation and the legislature’s democratic legitimacy is built into the text of the notwithstanding clause.

Second, the text adopts a solution to protecting core electoral rights and freedoms from the notwithstanding clause that is at best partial and incomplete. Many rights and freedoms essential to flourishing democratic competition are outside of sections 3-5 and are thus potentially subject to the notwithstanding clause. These most importantly include the freedom of association and expression in s.2 and the equality rights in s.15. In other words, the risk to democratic rights and freedoms remains considerable even if ss.3-5 are excluded from the scope of s.33.

a) The Exclusion of Sections 3-5 and Its Underlying Meaning

The exclusion of sections 3-5 from the ambit of s.33 is significant because these provisions protect core democratic rights. Section 3 enshrines the right of each citizen to vote ¹¹⁵ and to stand as a candidate ¹¹⁶ in federal and provincial elections. Section 3 also guarantees the “meaningful participation” ¹¹⁷ of individuals and their “effective representation” ¹¹⁸ when electoral boundaries are drawn. Unlike s.3, which pertains directly to the rights of individual citizens, sections 4 and 5 operate as structural protections for free and fair elections.

Section 4 sets five years as the maximum time between elections, as discussed in the section on electoral accountability. While legislation now often sets fixed election dates, s. 4(1) sets “an upper limit [between elections] to ensure regular accountability to the electorate” as a constitutional matter. ¹¹⁹ It works as an indirect guarantee of regular elections. ¹²⁰ The Court of Appeal for Alberta held that s. 4(1) codifies “the government’s right to call a general election at any time during the period of five years.” ¹²¹ There has been little litigation on s.4 because in the Charter era, legislatures have respected the five year maximum. ¹²²

Section 5 requires the House of Commons and provincial legislatures to meet at least once a year. ¹²³ It is based on earlier Canadian constitutional provisions and long-standing British tradition. One of the vulnerabilities of parliamentary democracy is that Parliament may simply not be convened on a regular basis. If it is not, then the executive may continue to act without having to test the confidence of the House. Particularly after an election that signals a likely change in government, a refusal to convene the legislature for an unreasonable amount of time undermines parliamentary democracy.

While not encompassing all of the minimum core of democracy, a universal franchise, regular and routine elections, and a requirement for the legislature to meet annually are necessary conditions for parliamentary democracy. ¹²⁴ All of the potential abuses by legislatures or governments of democratic rights that would have been made possible by an application of s.33 to ss.3-5 have long pedigrees in Canada, the U.K., and elsewhere. The framers anticipated Canada’s future on the basis of these past experiences. Prior to the Charter, for example, Parliament and the provincial legislatures repeatedly denied groups of citizens the right to cast a ballot, for partisan reasons or with discriminatory intent, against groups deemed unworthy. ¹²⁵

The text of the Charter therefore identifies misuse of s.33 in relation to democratic rights as one of the potential flaws in the new, post-1982 constitutional order that needed to be addressed in advance. The best interpretation of the exclusion of ‘democratic rights’ from the scope of s.33’s is an attempt to prevent legislatures or executives from undermining regular elections, transitions between governments, and a broad-based electorate of all adult citizens. The notwithstanding clause is structured in such a way as to attempt to minimize the chance that irresponsible legislative action could permanently harm the democratic foundations of the Canadian constitutional order.

b) The Problem of Scope

The exclusion of these ‘democratic rights’ as labelled in the Charter itself from s.33 is therefore an important and consequential decision and a recognition of the fragility of democracy. Sections 3-5 of the Charter, however, do not come close to covering all of the rights and freedoms directly relevant to ensuring a flourishing democracy. A host of other relevant Charter rights and freedoms to full realization of the ‘right of rights’ are still subject to s.33.

Freedom of political expression in s.2(b) ensures that individuals and groups can criticize the government and campaign for or against political parties and candidates. Section 2 also protects freedom of the press, which is undeniably important in a democracy. Pre-Charter, restrictions on the media or criticism of the government were sometimes imposed, especially by the Provinces. ¹²⁶ Freedom of association protects the capacity of individuals to organize themselves collectively into political parties, interest groups, ‘third party’ election advertisers, and so on. ¹²⁷ The equality right in s.15 guarantees that candidates and voters cannot be treated differentially on the basis of prohibited grounds in a manner that is discriminatory. There may be instances where s.15 rather than s.3 applies.

In many Charter cases involving political participation, it will be clear whether s.2, or s.3 or s.15 or some other provision is the applicable one. Where there is less clarity, the threshold question of which provision applies to a particular Charter claim ends up being consequential, as it determines whether s.33 can be invoked or not. ¹²⁸ This fact has always been the case since the introduction of the Charter in 1982, but was previously of marginal importance in the context of less regular use of s.33. The stakes are therefore high for categorizing a claim as either falling within or outside of the scope of s.33. While the text of s.33 prevents some of the worst potential outcomes for democracy, it still applies to rights and freedoms that have a significant role in ensuring a functioning electoral process.

V. Conclusion: Implications for Canadian Democracy

This article has argued that the prevailing theory of the legitimacy of the notwithstanding clause does not and, indeed cannot, establish all uses of s.33 by the legislature as ‘democratic’. Waldron’s commitment to democracy is an admirable one. Yet his theory cannot do the work that advocates for more frequent or regular use of s.33 in Canada claim that it does. Waldron’s account of democracy and judicial review is akin to a scientific theory that appears valid in the lab, but that fails under real world conditions. Waldron assumes away challenges that are central to the life of Canadian constitutional democracy.

Contrary to Waldron and his defenders, acknowledging the potential misuse of legislative power through s.33 does not entail faint-heartedness about democracy and does not compel us to adopt an elitist account that sneers at voters. We can be democrats with an optimistic account of the good intentions of most elected representatives, most of the time, while still recognizing the fallibility of legislatures and the structural barriers that operate against rights protection. We can respect the deliberative capacities and political judgement of voters without assuming that the electorate will always punish abuses of minority rights. We can be skeptical of the merits of the likely or potential uses of s.33 by legislatures without adopting the view that courts are infallible. Whether a particular decision by Parliament or a provincial legislature is democracy enhancing or detracting requires the political community to make such an assessment on a case-by-case basis.

Debates around the notwithstanding clause over its first 40 years have tended to focus on its general impact, rather than how it is used in relation to specific rights or freedoms. ¹²⁹ Particular attention to the relationship between s.33 and political rights and freedoms, however, is warranted in the current environment. The political restraint that allowed Canadians the luxury of not having to consider the full range of uses to which s.33 could be put appears to have dissolved. By placing the right to vote and some structural democratic rights outside of the scope of the notwithstanding clause, the text of s.33 prevents some exercises of the notwithstanding power in relation to democracy. There is still significant scope, however, for s.33 in relation to political participation. Given the potential consequences, special attention should be paid to the application of s.33 to political rights and freedoms.

If the basis for the notwithstanding clause is democracy, then on Waldron’s own account, invoking s.33 in a manner that harms the ‘right of rights’ should be seen as illegitimate. Equality of political participation is the foundation for his conclusion that the legislature possesses democratic legitimacy. If political participation is restricted, then the legislature’s legitimacy is reduced. While misuse of legislative authority so as to harm democracy has always been a theoretical concern with accounts that emphasize parliamentary sovereignty, such concerns are heightened in the context of more regular use of s.33. Unprincipled use of s.33 against the political rights and freedoms of unpopular minorities or political opponents would have serious, negative consequences for Canadian democracy.