Article Volume 60:4

Unsteady Architecture: Ambiguity, the Senate Reference, and the Future of Constitutional Amendment in Canada

This article critically examines the Supreme Court of Canada’s opinion in the Senate Reform Reference from the perspective of its coherence in interpreting the various amending procedures in Part V of the Constitution Act, 1982. It analyzes the ways that the underlying logic of the Court’s reasoning, particularly with respect to the method of selecting senators and senatorial term limits, creates ambiguity and risks unintended consequences for future attempts at constitutional amendment. The Court’s explicit refusal to distinguish between the federal government’s unilateral ability to enact a retirement age and its logic that term limits, regardless of length, require the consent of the provinces under the general amending procedure lacks logical consistency and arguably erodes the unilateral amending procedure to a problematic degree. In the context of its reasoning with respect to changes to the method of selecting senators, the Court’s reliance on the amorphous notion of the “constitutional architecture” clouds the definable limits of “method of selection” under section 42(1)(b). The Senate Reform Reference introduces considerable ambiguity into what changes the federal executive can implement with respect to the appointments process itself. The article concludes by exploring the political implications that the decision has for the future of Senate reform specifically and for our ability to amend the constitution generally.

Cet article porte un regard critique sur la décision de la Cour suprême du Canada dans le Renvoi relatif au Sénat. Il aborde sa cohérence selon la perspective de l’interprétation des procédures d’amendement contenues dans la Partie V de la Loi constitutionnelle, 1982. L’article analyse les moyens dont la logique sous-jacente au raisonnement de la Cour, particulièrement en ce qui a trait à la méthode de sélection des sénateurs et aux limites apportées à leurs mandats, crée de l’ambiguïté et court le risque de créer des conséquences involontaires affectant de futures tentatives d’amendements constitutionnels. La Cour refuse explicitement de distinguer entre la capacité qu’a le gouvernement fédéral d’imposer unilatéralement un âge de retraite et sa propre logique, selon laquelle la procédure générale d’amendement exige le consentement des provinces dans le cas de l’imposition de limites aux mandats des sénateurs, peu importe leur durée. Ceci démontre un manque de cohésion dans sa logique et érode la procédure d’amendement unilatérale à un degré problématique. D’ailleurs, le raisonnement de la Cour ayant trait au changement des modes de sélection des sénateurs se fie sur la notion amorphe de « l’architecture constitutionnelle » et, ce faisant, brouille les limites définissables de ce qui constitue un « mode de sélection » selon l’article 42(1)(b). Le Renvoi relatif au Sénat introduit un niveau considérable d’ambiguïté à l’égard des modifications du processus même de nomination sénatoriale que peut effectuer l’exécutif fédéral. Cet article conclut en explorant les implications politiques de la décision pour l’avenir de la réforme du Sénat, de façon spécifique, et pour notre capacité d’amender la constitution, de façon plus générale.

* Assistant Professor, Department of Political Science, University of Waterloo. My thanks to Alexander Pless and the McGill Law Journal’s anonymous reviewers for their thoughtful comments on earlier versions of this paper.

Table of Contents


I.  Critiquing the Senate Reference

A.  Consultative Elections

B.  Term Limits

C.  Senate Abolition

II. Implications of the Senate Reference



In the spring of 2014, the Supreme Court of Canada rendered decisions in the Supreme Court Act Reference[1] and the Senate Reference,[2] which together marked the Court’s first comprehensive examination of the constitutional amending formula in Part V of the Constitution Act, 1982.[3] This article critically examines the Senate Reference from the perspective of its coherence in interpreting the various amending procedures. It argues that the underlying logic of the Court’s reasoning, specifically with respect to the method of selecting senators and senatorial term limits, creates ambiguity and risks unintended consequences for future attempts at constitutional amendment generally and for Senate reform specifically.

The federal Conservative government had long sought to implement Senate reform through ordinary statute, a policy that has been a key feature of the Conservative Party’s electoral platforms since it first formed government in 2006.[4] The objective was to implement term limits for senators and consultative elections as part of the Senate appointments process. After repeated efforts to pass a bill,[5] the government finally acquiesced to critics who argued that the changes required formal constitutional amendment and submitted a set of reference questions to the Supreme Court for it to determine whether Parliament could enact such reforms without the approval of the provinces. The reference also asked which procedure under Part V was required for Senate abolition.[6] The constitutional formula under Part V includes five to seven different procedures for amendment, depending on how they are counted. Of particular relevance for the Senate Reference questions are: the general procedure, which requires the consent of Parliament and at least seven provinces representing at least fifty per cent of the population;[7] section 41 of the Constitution Act, 1982, which mandates unanimous consent of the provinces for particular changes, including changes to Part V itself;[8] section 42(1)(b), which specifically states that amendments pertaining to “the powers of the Senate and the method of selecting Senators” must be done under the general formula;[9] and section 44, which states that “[s]ubject to sections 41 and 42, Parliament may exclusively make laws amending the Constitution of Canada in relation to the executive government of Canada or the Senate and House of Commons.” Ultimately, the Court reached a unanimous opinion that provincial consent for term limits and consultative elections is required under the general amending procedure and that unanimity is required for abolition.

The Court’s approach to assessing the constitutional amending procedures begins with a starting point, articulated in the Secession Reference,[10] that constitutional interpretation involves examining “the constitutional text itself, the historical context, and previous judicial interpretations of constitutional meaning.”[11] The justices note as well that “constitutional interpretation must be informed by the foundational principles of the Constitution, which include principles such as federalism, democracy, the protection of minorities, as well as constitutionalism and the rule of law.”[12] It is through these principles that the Court concludes, as it has invoked in earlier cases,[13] that the Constitution ought to be regarded as having an “internal architecture”[14] or “basic constitutional structure,”[15] meaning that the constitution

must be interpreted with a view to discerning the structure of government that it seeks to implement. The assumptions that underlie the text and the manner in which the constitutional provisions are intended to interact with one another must inform our interpretation, understanding, and application of the text.[16]

Describing the Constitution’s architecture is in line with a purposive approach to interpretation that seeks to capture the meaning of specific constitutional provisions and to prevent interpretations that conflict with or contradict the application of other components of the constitution. It also underscores, as the justices point out, that amendments to the constitution are not limited to textual changes, but also apply to changes to the way the constitution operates. Therefore, on the one hand, an appreciation of the constitutional architecture ensures specific provisions are interpreted to operate as parts of a coherent whole.

On the other hand, however, too much dependence on the fundamentally vague notion of the basic structure of the constitution may divorce specific provisions from their textual underpinnings and their basic meaning. A reliance on the concept of the constitution’s architecture also gives the justices considerable discretion in choosing how to locate and define specific issues depending on how they view the broader governing structure. Interpreting specific constitutional provisions with too much of a focus on the indeterminate constitutional structure rather than rooting analysis more directly in the text thus risks a great level of dependence on the justices’ ability to accurately describe the various institutions, conventions, and processes that animate the constitution.

In what follows, I argue that the Court relies too heavily on the concept of constitutional architecture in its reasons when a slightly more narrow, more textually rooted approach would have been sufficient to arrive at a coherent dividing line between the various amending procedures and to establish a clear standard for future assessments of which procedures are required for changes relating to the Senate. Further, where the justices tread too far in exploring aspects of the constitutional architecture, they do not go far enough in examining the amending formula’s specific provisions, such as section 44 of Part V of the Constitution Act, 1982, where they fail to provide a logical justification for the minimal role they outline for Parliament in effecting changes to the Senate. This article elaborates on this critique of the justices’ logic relating to changes that would enact consultative elections and term limits for senators, while also briefly explaining why the Court arrived at the correct conclusion with respect to the abolition of the Senate. The remainder of the article then examines the implications of the Court’s reasoning.

I.  Critiquing the Senate Reference

A.  Consultative Elections

With respect to consultative elections, the federal government posed the following questions to the Court:

  1. Is it within the legislative authority of the Parliament of Canada, acting pursuant tosection 91 of the Constitution Act, 1867, or section 44  of the Constitution Act, 1982, to enact legislation that provides a means of consulting the population of each province and territory as to its preferences for potential nominees for appointment to the Senate pursuant to a national process as was set out in Bill C-20, the Senate Appointment Consultations Act?
  2. Is it within the legislative authority of the Parliament of Canada, acting pursuant tosection 91 of the Constitution Act, 1867, or section 44  of the Constitution Act, 1982, to establish a framework setting out a basis for provincial and territorial legislatures to enact legislation to consult their population as to their preferences for potential nominees for appointment to the Senate as set out in the schedule to Bill C-7, the Senate Reform Act?[17]

Under section 42(1)(b) of the Constitution Act, 1982, changes to “the powers of the Senate and the method of selecting Senators” must be made according to the general amending procedure. The federal government argued that the prime minister would retain full discretion to make the final decision on senatorial appointments under a system of strictly advisory elections, and therefore their implementation would not constitute a change to the selection process. The justices did not give much credence to this argument, noting that while in theory the prime minister might refuse to make appointments based on electoral outcomes, the very purpose of the reforms the government sought was “to bring about a Senate with a popular mandate.”[18] They elaborate: “[w]e cannot assume that future prime ministers will defeat this purpose by ignoring the results of costly and hard-fought consultative elections.”[19] The justices thus state that the federal government’s argument incorrectly privileges “form over substance” in its interpretation of the meaning of 42(1)(b).[20] If advisory elections are advisory in name only, then their implementation effectively provides a loophole to escape the requirements of the general amending procedure.

It is worth noting that this finding privileges a particular conception of “method of selection.” The narrow reading espoused by the federal government views the executive’s final decision making authority as the central element. From a certain perspective, there is a logical coherence to this view. Up until now, the actual process that precedes the formal recommendation of the prime minister and appointment by the Governor General has been at the virtually unfettered discretion of the prime minister, who has been free to canvass and consult anyone for names to consider. Candidates are routinely selected on the basis of patronage and often as a result of past work of a partisan nature, but in theory, the prime minister has been free to make the final determination as the result of a committee of staffers in the Prime Minister’s Office, the recommendation of cabinet colleagues, or even the flip of a coin. In a narrow respect, the Court’s decision arguably creates an absurdity: a prime minister is free to consult with whomever he or she wishes except for the voting public.

Yet the Court’s reasoning also makes clear that the prime minister is not free to bind his or her decision making authority in all practical senses. It is difficult to dispute the justices on this point: elections, even if technically consultative, come with the baggage of democratic legitimacy that makes it very difficult to foresee the appointment of winners of senatorial election campaigns becoming anything other than normal practice. As a result, advisory elections would mark a significant change in the way the final, formal decision to appoint is made, even if by conventional practice and not as a matter of formal law. In the result, the Court’s reasoning here provides at least a legitimate legal grounding for interpreting section 42(1)(b) so that it applies to the institution of advisory elections.

The justices, however, extend their rationale for this conclusion beyond the scope of a contextual analysis of how section 42(1)(b) applies with respect to establishing advisory elections. In line with its emphasis on the constitutional architecture concept, the Court describes the impact consultative elections would have on the operation of the Senate as an institution. The justices point out that the

framers sought to endow the Senate with independence from the electoral process to which members of the House of Commons were subject, in order to remove Senators from a partisan political arena that required unremitting consideration of short-term political objectives.[21]

The justices argue

the choice of executive appointment for Senators was also intended to ensure that the Senate would be a complementary legislative body, rather than a perennial rival of the House of Commons … This would ensure that they would confine themselves to their role as a body mainly conducting legislative review, rather than as a coequal of the House of Commons.[22]

This conception of the Senate “shapes the architecture”[23] of the Constitution Act, 1867.[24] Advisory elections would constitute a significant change to the function of the Senate and, as a federal institution, such a change requires provincial consent.

A number of problems flow from the Court’s foray into describing the Senate’s function. First, the justices’ description of the Senate as a body of “sober second thought” and one that engages primarily in “legislative review” is somewhat simplistic. Studies of the upper house show that the classic sober second thought depiction is overstated, in part because the Senate has more legislative influence than is normally recognized.[25] As David E. Smith notes, the Senate has a major role to play in constraining majority governments.[26] It has even, for a time, been involved in scrutinizing draft bills before their introduction in the House (pre-study), which had an important impact on legislative drafting. None of this is to suggest the Senate has been a competitive rather than a complementary body, but there are important nuances to the Senate’s function that the Court’s depiction does not fully address.

Second, the Court’s emphasis on the proposed reform’s impact—specifically, giving the upper house a “democratic mandate”—belies other potential changes that might affect the Senate’s role vis-à-vis the House of Commons. It is not surprising that the Court left unaddressed what other reforms to the appointments process might be permissible without requiring the general amending procedure. The justices were addressing the questions before them and wading too deeply into hypothetical scenarios would be fraught with difficulty (and is also generally avoided in the context of references). Nonetheless, by not resting their reasons on the determination that the particular electoral reform proposal falls under the ambit of section 42(1)(b), the justices clouded the issue of whether other, more modest reforms to the process are possible when they invoked the Senate’s general attitude as a deferential body of sober second thought. Nor does the Court’s opinion settle solely on the fact that the proposed process would, in practice, be binding on the prime minister’s discretion; it also emphasized that the specific nature of elections—and the democratic mandate they give the Senate itself—were particularly likely to alter the Senate’s function.[27]

The Court’s opinion not only fails to provide guidelines about what changes Parliament might make to rules governing senatorial selection or even a consultative process that might draw up a list of potential candidates for appointment, but it also arguably clouds the issue more than if it had simply rested its reasons on its discussion of section 42(1)(b). This problem is more than hypothetical because, by the time the Court rendered its opinion, there were already other proposals put forward in political debate.[28] Liberal leader Justin Trudeau has proposed abolishing partisanship and patronage as factors in the senatorial selection process.[29] A non-partisan appointments process would mark a departure from past practice, and in the extreme could, over time, remove partisanship as a feature from the Senate entirely. Moreover, the Liberals have promised an “open, transparent and non-partisan appointment process for Senators … informed by other non-partisan appointment processes, such as that of the Supreme Court Justices and Order of Canada recipients.”[30]

While the specific details of the proposal remain to be seen, the prospect of an arm’s-length committee, composed of non-partisan members who develop a short list of candidates (or even propose nominees) for the prime minister, raises a number of interesting questions in light of the Court’s reasons. The Court’s opinion suggests that this may require an amendment under the general procedure, depending on how much emphasis is given to the electoral nature of the impugned proposals, when it writes

[t]he words “the method of selecting Senators” include more than the formal appointment of Senators by the Governor General. … The proposed consultative elections would produce lists of candidates, from which prime ministers would be expected to choose when making appointments to the Senate. The compilation of these lists through national or provincial and territorial elections and the Prime Minister’s consideration of them prior to making recommendations to the Governor General would form part of the “method of selecting Senators”. Consequently, the implementation of consultative elections falls within the scope of s. 42(1)(b) and is subject to the general amending procedure.[31]

It is not clear whether this applies to any list produced by any process from which the prime minister would be “expected” to choose names when making appointments, or if there is something particular to a list drawn from an electoral process. If the former, then the Court’s opinion may place much greater restrictions on the front end of the selection process than many observers previously contemplated.

From the perspective of the Senate’s function, a merit-based process that removes partisanship and patronage from the appointments process would also confer added legitimacy to the Senate as a body. While not necessarily of the same magnitude or character as the democratic mandate afforded by an electoral process, a non-partisan Senate composed of eminent Canadians appointed from a quasi-independent process would arguably mark a fundamental shift in the nature of the Senate’s composition. It is conceivable that senators appointed via a non-partisan process, and whom the public came to view with more respect as a result of not being beneficiaries of patronage, would recognize, and be emboldened by, the added perceived legitimacy such a context afforded. Although unlikely to transform the Senate into a body that would act in constant competition with the House of Commons, such a change could nonetheless lead to more frequent legislative activity in the form of amendments and even vetoes to bills coming up from the lower house. The Court’s reference opinion leaves much doubt about what degree of change in the Senate’s function is sufficient to require a constitutional amendment under the general formula.

It is also unclear to what extent the prime minister might introduce new elements—even informal ones—into the process leading up to the final selection of senators that might bind his or her discretion. There is likely an important distinction to be made between an effort to formally establish a process in law (and thereby attempt to bind future prime ministers to a particular process) versus informally constituting a committee or some new process of consultation before making an appointment. This raises a fundamental question of whether these sorts of distinctions amount to constitutional hairsplitting. The notion that a prime minister can implement certain reforms, but only if he or she does so informally, speaks to the flexibility of our constitutional architecture, perhaps, or it may simply expose a fundamental logical inconsistency within the Court’s interpretation of the amending formula. Is a prime minister free to attempt, informally, to establish a convention of appointing senators of a certain type to the Senate or by some particular process by which candidate names are produced? The Court’s opinion raises the question but does not answer it.

This question relates to another issue that is left unclear by the Court’s opinion: whether provinces are still free to run their own Senate elections. The reference asked whether Parliament could pass legislation to run its own elections or “establish a framework” for provinces to consult their electors.[32] Yet Alberta has long held Senate elections on its own initiative; indeed, a number of senators who have won these contests were subsequently appointed to the upper chamber. The Court did not comment on this more informal (from the federal government’s perspective) process, or on the legitimacy of these particular senators’ standing in the Senate. From a constitutional perspective, it would be surprising if a province were somehow prohibited from canvassing its voters via plebiscite on any matter it wished. It is not clear whether the prime minister is now prohibited from exercising the discretion to appoint someone who had won one of these provincially administered contests.[33]

B.  Term Limits

On term limits, the Court was asked the following:

  1. In relation to each of the following proposed limits to the tenure of Senators, is it within the legislative authority of the Parliament of Canada, acting pursuant tosection 44 of the Constitution Act, 1982, to make amendments to section 29  of the Constitution Act, 1867 providing for

(a) a fixed term of nine years for Senators, as set out in clause 5 of Bill C-7, the Senate Reform Act;

(b) a fixed term of ten years or more for Senators;

(c) a fixed term of eight years or less for Senators;

(d) a fixed term of the life of two or three Parliaments for Senators;

(e) a renewable term for Senators, as set out in clause 2 of Bill S-4, Constitution Act, 2006 (Senate tenure);

(f) limits to the terms for Senators appointed after October 14, 2008 as set out in subclause 4(1) of Bill C-7, the Senate Reform Act; and

(g) retrospective limits to the terms for Senators appointed before October 14, 2008?[34]

On the question of term limits, the Court confronted arguably the most difficult issue of the reference. Unlike “the method of selecting senators,” the specific issue of senatorial terms is not explicitly mentioned in Part V.[35] At issue with respect to term limits was whether Parliament could implement them unilaterally under section 44 or if provincial consent was required under the general amending procedure. The justices concluded that, as with consultative elections, term limits could only be implemented under the general amending formula.

While noting that senatorial terms were not an issue encompassed by changes referred to in section 42(1)(b), the Court stated that provinces have an interest in any changes affecting the “fundamental nature or role” of the Senate.[36] Specifically, the justices write, “it does not follow that all changes to the Senate that fall outside of s. 42 come within the scope of the unilateral federal amending procedure in s. 44.”[37]

The Court’s approach to interpreting section 44 is guided by its understanding of the historical context surrounding that provision. In 1949, the British North America Act (No. 2), 1949[38] inserted section 91(1) into the Constitution Act, 1867 and gave the Canadian Parliament broad new authority over constitutional amendments, with the exception of anything relating to provincial powers or rights, minority education, and languages. Prior to this, only a minimal set of changes to the original Constitution Act, 1867 could be made domestically, as permitted through a handful of specific provisions. Parliament was authorized to make “housekeeping” changes to the Senate or House of Commons, such as increasing the number of members of Parliament (under section 52), establishing and changing electoral districts (section 40), changing quorum in the Senate (section 35) and amending the privileges and immunities of Members of Parliaments (section 18), and allowing provinces to make changes to provincial constitutions, so long as these did not affect the lieutenant governor (section 92(1)).[39] The 1949 changes broadening the scope of federal amending authority were made without provincial consent:

The federal position was that provincial consent was unnecessary because the new amending power was of concern to the federal government alone and could not be used to affect provincial powers. The provinces rejected this justification and claimed that section 91(1) could nonetheless operate to permit the federal government to enact amendments that would indirectly affect provincial interests in the federation.[40]

In the Senate Reference, the Court interpreted section 44 as effectively a replacement for this old provision, and determines that, despite the broad textual language of section 44, Parliament can only effect housekeeping changes to the Senate. Section 44 is thus regarded as a circumscribed exception to the general amending procedure.[41]

Any changes that might alter the Senate’s function or role, or have any implications for the provinces, require provincial consent. In the context of term limits, the justices state:

[T]he Senate’s fundamental nature and role is that of a complementary legislative body of sober second thought. The current duration of senatorial terms is directly linked to this conception of the Senate. Senators are appointed roughly for the duration of their active professional lives. This security of tenure is intended to allow Senators to function with independence in conducting legislative review. This Court stated in the Upper House Reference that, “[a]t some point, a reduction of the term of office might impair the functioning of the Senate in providing what Sir John A. Macdonald described as ‘the sober second thought in legislation’”: p. 76. A significant change to senatorial tenure would thus affect the Senate’s fundamental nature and role.[42]

It is notable that the Court acknowledges its statement in the 1980 Upper House Reference, where it held that the imposition of mandatory retirement age “did not change the essential character of the Senate”[43] and can, therefore, be regarded as a legitimate exercise of unilateral action by Parliament. Constitutional scholars who have noted that the procedures of Part V as they relate to the Senate were an attempt to “codify” the Court’s opinion in the 1980 reference have also argued that it would be acceptable for Parliament to enact term limits of a certain length under section 44:[44]

The items specified in section 42 should be regarded as an exhaustive list of matters deemed fundamental or essential, as those terms were utilized in the Senate Reference. To hold that the unilateral federal power in section 44 is subject to a further limitation along the lines suggested would lead to needless uncertainty and ambiguity.[45]

Importantly, the Court explicitly refused to address the seemingly pertinent question of why a retirement age might fall under the category of “housekeeping” but the imposition of terms limits of any length or design do not. The justices write that

[i]t may be possible, as the Attorney General of Canada suggests, to devise a fixed term so lengthy that it provides a security of tenure which is functionally equivalent to that provided by life tenure. However, it is difficult to objectively identify the precise term duration that guarantees an equivalent degree of security of tenure.[46]

The Court’s refusal to engage in a line-drawing exercise here is problematic to the extent that line-drawing is precisely what was being asked of it. The structure of the federal government’s reference questions on term limits was clearly designed to encapsulate a range of alternatives. The questions posed to the Court provide a clear indication that guidance was sought as to whether certain types of term limits might be enacted under section 44 even if other types could not. By not addressing the question of whether a non-renewable term limit long enough to avoid altering the basic function of the Senate or the role of senators is feasible, the Court sidestepped a contradiction and logical flaw in its approach to interpreting section 44. Presuming the validity of the Court’s own interpretation of section 44’s development, there is no reason to believe that Parliament would not be theoretically free to unilaterally lower the mandatory retirement age of senators to seventy or sixty-five. However, it is not free, according to the Court, to unilaterally enact a non-renewable term limit of fifteen years. The idea that one of these changes would alter the fundamental features of the institution and the other would not is difficult to comprehend. By refusing to engage with this question, the Court has not only failed to deliver a good standard by which some matters may fall under different procedures in Part V, but it has also arguably gutted section 44 in a way not reasonably contemplated.

The justices’ decision to refuse to distinguish between different types of term limits is also contrary to existing evidence that not all term limits would alter how the Senate functions. Fixed terms would undoubtedly run the risk of impairing the Senate’s independence if they were renewable. Similarly, excessively short term limits might make a Senate appointment a brief mid-career stint, something that might alter senators’ approach to their role and skew their decision-making incentives. But there is little evidence that lengthy, non-renewable terms would pose similar dangers. As Christopher Manfredi writes in his expert submission for the reference, the “average age at which individuals have been appointed is 57, which means that, had the nine-year fixed term applied since 1867, the average senator would have left the Senate at age 66 and would not have expected a lengthy post-Senate career.”[47] Moreover, the mean and median length of senatorial service since 1965 has been 11.3 and 9.8 years, respectively.[48] The imposition of lengthy non-renewable terms—be they nine, twelve, or fifteen years—would not constitute a departure from the Senate’s existing reality, nor could it realistically be thought to alter the Senate’s fundamental features or operation.

C.  Senate Abolition

On the issue of abolishing the Senate, the Court was asked the following questions:

  1. Can an amendment to the Constitution of Canada to abolish the Senate be accomplished by the general amending procedure set out insection 38 of the Constitution Act, 1982, by one of the following methods:

(a) by inserting a separate provision stating that the Senate is to be abolished as of a certain date, as an amendment to the Constitution Act, 1867  or as a separate provision that is outside of the Constitution Acts, 1867 to 1982 but that is still part of the Constitution of Canada;

(b) by amending or repealing some or all of the references to the Senate in the Constitution of Canada; or

(c) by abolishing the powers of the Senate and eliminating the representation of provinces pursuant to paragraphs 42(1)(b) and (c) of the Constitution Act, 1982?

  1. If the general amending procedure set out insection 38 of the Constitution Act, 1982  is not sufficient to abolish the Senate, does the unanimous consent procedure set out in section 41  of the Constitution Act, 1982 apply?[49]

Although the question on abolition was perhaps the most straightforward issue in the reference for the Court, a number of odd or erroneous arguments arose during political debate over the issue. For example, Robert Ghiz, then-premier of Prince Edward Island, spoke out against Senate abolition because it would mean his province “would be down to one member of Parliament” as a result of the constitutional guarantee giving Prince Edward Island the same number of Members of Parliament as it has senators.[50] The premier’s fears are unfounded, however, because section 41(b) of the amending formula preserves the number of members in the House of Commons for each province such that they do not fall below their number of senators “at the time this Part comes into force.”[51] In other words, the abolition of the Senate would not reduce the number of Members of Parliament to which Prince Edward Island is entitled barring a unanimous amendment that alters section 41(b).

With respect to Senate abolition, any change to the amending formula itself requires unanimity under section 41(e), and the Senate is referenced throughout Part V. The federal government attempted to argue that the Senate could be abolished under the general amending formula without amending the text of Part V, as references to the Senate in Part V would be viewed as spent provisions following any general amendment to do so. Notably, section 47 of Part V provides the Senate with a suspensive veto that requires the House of Commons to adopt a second resolution after 180 days if the Senate refuses to adopt an initial resolution to amend the constitution under any of the procedures other than sections 44 or 45.[52] The federal government argued that, because the Senate could be overridden after 180 days under section 47, references to it in the amending formula were incidental to its abolition. It is worth noting that such logic could be flipped on its head: the fact that the Senate itself was granted the power to delay amendments for 180 days only underscores its relevance and the significance of its presence in Part V. The ratification of resolutions to amend the constitution when provincial consent is required is a difficult process, and a 180-day delay could result in an intervening election in some provinces. During the ratification process for the Meech Lake Accord,[53] the election of the Clyde Wells government resulted in Newfoundland and Labrador revoking its assent to the Accord and contributed to the failure of the constitutional package.[54] For this reason, the Senate’s suspensive veto should be regarded as having substantive, in addition to symbolic, significance.

For its part, the Court concluded correctly that

Part V was drafted on the assumption that the federal Parliament would remain bicameral in nature, i.e. that there would continue to be both a lower legislative chamber and a complementary upper chamber. Removal of the upper chamber from our Constitution would alter the structure and functioning of Part V. Consequently, it requires the unanimous consent of Parliament and of all the provinces (s. 41(e)).[55]

II. Implications of the Senate Reference

The rationale employed by the Court in the Senate Reference raises a number of issues with respect to Senate reform specifically and future constitutional amendment generally. In the context of their reasoning regarding changes to the method of selecting senators, the justices’ reliance on the amorphous notion of the constitutional architecture clouds the definable limits of method of selection under section 42(1)(b). While it would be unreasonable to expect the reference opinion to account for every hypothetical reform option, the Court should be counted on to provide clear guidelines about the scope of the relevant provisions in Part V. Instead, the reasoning adopted by the justices adds to, rather than alleviates, the uncertainty over what other changes to the appointments process might be feasible without formal constitutional amendment and provincial consent. This is not to argue that the Court reached an incorrect decision on whether consultative elections fall under the ambit of section 42(1)(b); nor is it to suggest that only a narrow, textual reading of the constitution is the appropriate jurisprudential approach. However, the appeal to the constitution’s broader architecture introduces ambiguity where a focus on a contextual reading of 42(1)(b) would have sufficed.

By contrast, the Court’s appeal to constitutional architecture not only leads it to introduce problematic uncertainty into its discussion of the imposition of term limits, it also leads the justices to an incorrect conclusion regarding Parliament’s ability to enact changes unilaterally under section 44. The Court’s explicit refusal to distinguish between the federal government’s ability to enact a retirement age under section 44, and its logic that term limits, regardless of length, require the consent of the provinces under the general amending procedure lacks logical consistency and arguably erodes section 44 to a problematic degree. The justices’ rationale also flies in the face of available evidence about the conceivable impact lengthy, non-renewable terms limits would have on the Senate’s function and essential features. The structure of the questions on term limits posed to the Court makes it clear that the justices were being asked to examine whether some types of term limits might avoid implicating the general amending procedure even if others do. By sidestepping this question and explicitly refusing to engage the question of whether lengthy term limits might not alter the Senate’s function, the Court effectively avoids dealing with the central issue at stake. The effect of this is to minimize section 44 in a manner not necessarily compatible with its historical context or even with the Court’s stated rationale.

The Court’s reasoning may also have the effect of producing unintended (or at least unanticipated) consequences, assuming its logic is employed consistently for other issues implicating the amending formula. One of the most obvious examples of unintended consequences emanates from the Court’s opinion in the Supreme Court Act Reference. In that opinion, a majority of justices determined that the eligibility requirements for justices of the Supreme Court of Canada as outlined in sections 5 and 6 of the Supreme Court Act are entrenched in the Constitution.[56] Any changes to the eligibility requirements thus require unanimous approval of the provinces because they fall under section 41(d) of Part V as part of the “composition of the Supreme Court of Canada.”[57] Prior to the Court’s decision, adding additional eligibility requirements under the Supreme Court Act, such as mandating bilingualism for Supreme Court appointees—a policy supported by the Liberal Party and New Democratic Party—was not thought to have required constitutional amendment. This decision now renders any such changes extremely unlikely, as there has never been an amendment successfully ratified under the constitution’s unanimity procedure and the current political climate is such that this is unlikely to change.[58]

As it relates to the Senate Reference, the Court’s appeal to constitutional architecture, together with its general antipathy toward indirect methods of amending the constitution, has obvious potential implications for other, even informal, changes to the senatorial appointments process, as discussed above. Moreover, the Court’s rationale might have implications for statutes like An Act Respecting Constitutional Amendments.[59] Following the 1995 Quebec secession referendum, the federal government enacted An Act Respecting Constitutional Amendments to effectively provide Quebec (and, by design, Canada’s other regions) a veto over most major amendments. The Act prohibits government ministers from proposing constitutional resolutions unless consent is first obtained from Ontario, Quebec, British Columbia, at least two Atlantic provinces representing at least fifty per cent of the Atlantic populations, and at least two prairie provinces representing at least fifty per cent of the prairie population (in effect, giving Alberta its own veto).[60] The Act effectively uses the federal government’s inherent veto under most of Part V’s amending procedures to establish a system of regional vetoes for constitutional amendment.[61] The Act was passed to fulfill the government’s commitment to Quebec federalists but, by giving Ontario, British Columbia, and (in effect) Alberta a veto, it makes future constitutional reform—including reform desired by Quebec federalists—considerably more difficult.[62]

The Court’s general rationale regarding the ability of Parliament to implement changes to the constitution without the consent of the provinces would seem to apply to An Act Respecting Constitutional Amendments, which in practice acts as a unilateral amendment to Part V. From the perspective of the basic architecture of the constitution, the law compels the federal government to restrict the legitimate exercise of Part V without the consent of provinces. Moreover, it goes beyond a political decision by a government about whether to support a resolution because it binds future governments in law. If Parliament is not free to implement changes to the method of selecting senators or to impose term limits without provincial consent, it would be inconsistent with the Court’s approach to interpreting Part V to permit such disregard for the basic constitutional architecture, which includes the foundational agreement about the structure of the amending procedures in Part V.

From a broader perspective, the Court’s approach to interpreting the amending formula might also be evaluated on the basis of the balance between flexibility and rigidity that animates Part V’s design. If the Constitution represents the fundamental rules and structure for the country’s governing system, then the amending formula determines who gets to write those rules. A good formula provides enough flexibility for change to occur when it is needed or where there is sufficiently deep and broad consensus, but ensures enough rigidity so that fundamental changes can only be accomplished where consensus warrants. Debate during constitutional negotiations over the design of Part V recognized this tension.[63] The 1982 agreement resulted in a complex set of procedures designed to accommodate the need for consensus over fundamental change while ensuring the flexibility to prevent constitutional stasis. Judicial interpretation of these various procedures that expands the application of some over others risks imbalance, the result of which is to either lower the threshold for constitutional change to the point of disregarding the need for consensus or raising it so high as to invite constitutional stasis.


On this score, the Court’s Senate Reference opinion is problematic for two reasons. First, as noted above, the justices minimize section 44, characterizing it as a very narrow exception to the general amending procedure. As I have argued, they do so without sufficiently addressing the logical inconsistency of their approach. Second, the Court’s appeal to the basic structure or architecture of the constitution ultimately obscures rather than clarifies the dividing line between the various amending procedures. This may have the effect of chilling future attempts at constitutional change or, at the very least, increasing contestation and future legal challenges to reform efforts. Greater clarity through more coherent guidelines about the scope of the various amending procedures would have reduced this uncertainty. In that respect, the Senate Reference is a failure.


[1]     Reference Re Supreme Court Act, ss 5 and 6, 2014 SCC 21, [2014] 1 SCR 433 [Supreme Court Act Reference].

[2]     Reference Re Senate Reform, 2014 SCC 32, [2014] 1 SCR 74 [Senate Reference].

[3]     Being Schedule B to the Canada Act 1982 (UK), 1982, c 11 [Constitution Act, 1982].

[4]     See Conservative Party of Canada, Stand Up for Canada (2006); Conservative Party of Canada, The True North Strong and Free: Stephen Harper’s Plan for Canadians (2008); Conservative Party of Canada, Here for Canada: Stephen Harper’s Low-Tax Plan for Jobs and Economic Growth (2011).

[5]     See Bill C-7, An Act Respecting the Selection of Senators and Amending the Constitution Act, 1867 in Respect of Senate Term Limits, 1st Sess, 41st Parl, 2011 (first reading 21 June 2011); Bill C-10, An Act to Amend the Constitution Act, 1867 (Senate term limits), 3rd Sess, 40th Parl, 2010 (first reading 29 March 2010); Bill C-20, An Act to Provide for Consultations with Electors on their Preferences for Appointments to the Senate, 2nd Sess, 39th Parl, 2007 (first reading 13 November 2007); Bill C-19, An Act to Amend the Constitution Act, 1867, (Senate tenure), 2nd Sess, 39th Parl, 2007 (first reading 13 November 2007); Bill C-43, An Act to Provide Consultations with Electors on their Preferences for Appointments to the Senate, 1st Sess, 39th Parl, 2006 (first reading 13 December 2006).

[6]     The Court was also asked to address which amending procedure applied to the comparatively minor question of abolishing property requirements for senators.

[7]   Constitution Act, 1982, supra note 3, s 38(1):

An amendment to the Constitution of Canada may be made by proclamation issued by the Governor General under the Great Seal of Canada where so authorized by

(a) resolutions of the Senate and House of Commons; and

(b) resolutions of the legislative assemblies of at least two-thirds of the provinces that have, in the aggregate, according to the then latest general census, at least fifty per cent of the population of all the provinces.

[8]     See ibid, s 41:

An amendment to the Constitution of Canada in relation to the following matters may be made by proclamation issued by the Governor General under the Great Seal of Canada only where authorized by resolutions of the Senate and House of Commons and of the legislative assembly of each province:

(a) the office of the Queen, the Governor General and the Lieutenant Governor of a province;

(b) the right of a province to a number of members in the House of Commons not less than the number of Senators by which the province is entitled to be represented at the time this Part comes into force;

(c) subject to section 43, the use of the English or the French language;

(d) the composition of the Supreme Court of Canada; and

(e) an amendment to this Part.

[9]     Ibid, s 42(1), which lists specific amendments to the Constitution that may be made only in accordance with subsection 38(1):

(a) the principle of proportionate representation of the provinces in the House of Commons prescribed by the Constitution of Canada;

(b) the powers of the Senate and the method of selecting Senators;

(c) the number of members by which a province is entitled to be represented in the Senate and the residence qualifications of Senators;

(d) subject to paragraph 41(d), the Supreme Court of Canada;

(e) the extension of existing provinces into the territories; and

(f) notwithstanding any other law or practice, the establishment of new provinces.

[10]    Reference Re Secession of Quebec, [1998] 2 SCR 217, 161 DLR (4th) 385 [Secession Reference].

[11]    Senate Reform Reference, supra note 2 at para 25, citing Secession Reference, supra note 10 at para 32.

[12]    Ibid.

[13]    See Secession Reference, supra note 10 at para 50; OPSEU v Ontario (AG), [1987] 2 SCR 2 at 57, 41 DLR (4th) 1 [OPSEU].

[14]    Secession Reference, supra note 10 at para 50.

[15]    OPSEU, supra note 13 at 57.

[16]    Senate Reform Reference, supra note 2 at para 26.

[17]    Senate Reform Reference, supra note 2 at para 5.

[18]    Ibid at para 62.

[19]    Ibid.

[20]    Ibid at para 52.

[21]    Ibid at para 57.

[22]    Ibid at para 58 [emphasis in original].

[23]    Ibid at para 59.

[24]    (UK), 30 & 31 Vict, c 3, reprinted in RSC 1985, Appendix II, No 5 [Constitution Act, 1867].

[25]    See Paul G Thomas, “Comparing the Lawmaking Roles of the Senate and the House of Commons” in Serge Joyal, ed, Protecting Canadian Democracy: The Senate You Never Knew (Montréal: McGill-Queen’s University Press, 2003) 189 at 189; David E Smith, The Canadian Senate in Bicameral Perspective (Toronto: University of Toronto Press, 2003) at 114.

[26]    See ibid.

[27]    See Senate Reference, supra note 2 at para 52.

[28]    See e.g. Aaron Wherry, “Justin Trudeau’s Unilateral Senate Reform”, Maclean’s (29 January 2014), online: <>.

[29]    In the interests of full disclosure: I was asked to advise the Liberal Party on the constitutionality of their proposals prior to the release of the Court’s Senate Reference opinion.

[30]    See Wherry, supra note 28.

[31]    Senate Reference, supra note 2 at para 65.

[32]    Senate Reform Reference, supra note 2 at para 5.

[33]    Plans by the Government of New Brunswick to hold Senate elections were described as being in “limbo” following the Court’s decision. See Jacques Poitras, “David Alward’s Senate Reform Plans in Legal Limbo”, CBC News (30 April 2014), online: <>.

[34]    Senate Reference, supra note 2 at para 5.

[35]    Nonetheless, it was clear that an amendment would be required given senatorial tenure is listed in section 29(2) of the Constitution Act, 1867, which states that “[a] Senator who is summoned to the Senate … shall … hold his place in the Senate until he attains the age of seventy-five years.”

[36]    Senate Reference, supra note 2 at para 78.

[37]    Ibid at para 74.

[38]    (UK), 13 Geo VI, c 81, s 1.

[39]    See Patrick J Monahan & Byron Shaw, Constitutional Law, 4th ed (Toronto: Irwin Law, 2013) at 169–73.

[40]    Ibid at 171.

[41]    See Senate Reference, supra note 2 at para 75.

[42]    Ibid at para 79, citing Re: Authority of Parliament in Relation to the Upper House, [1980] 1 SCR 54 at 76, 102 DLR (3d) 1 [Upper House Reference].

[43]    Supra note 42 at 77.

[44]    Monahan & Shaw, supra note 39 at 213.

[45]    Ibid at 213–14.

[46]    Senate Reference, supra note 2 at para 81.

[47]    Christopher P Manfredi, An Expert Opinion on the Possible Effects of Bill C-7 (June 2013) at para 50.

[48]    See ibid at para 34.

[49]    Senate Reference, supra note 2 at para 5.

[50]    Teresa Wright, “Scrapping the Senate Bad for P.E.I., Local Politicians Argue”, The Guardian (28 May 2013), online: <,-local-politicians-argue/1>.

[51]    Constitution Act, 1982, supra note 3, s 41(b).

[52]    See ibid, s 47(1):

An amendment to the Constitution of Canada made by proclamation under section 38, 41, 42 or 43 may be made without a resolution of the Senate authorizing the issue of the proclamation if, within one hundred and eighty days after the adoption by the House of Commons of a resolution authorizing its issue, the Senate has not adopted such a resolution and if, at any time after the expiration of that period, the House of Commons again adopts the resolution and if, at any time after the expiration of that period, the House of Commons again adopts the resolution.

(2) Any period when Parliament is prorogued or dissolved shall not be counted in computing the one hundred and eighty day period referred to in subsection (1).

[53]    1987 Constitutional Accord (Ottawa: Library of Parliament, 3 June 1987) [Meech Lake Accord].

[54]    See Peter H Russell, Constitutional Odyssey: Can Canadians Become a Sovereign People?, 3rd ed (Toronto: University of Toronto Press, 2004) at 142–48.

[55]    Senate Reference, supra note 2 at para 106.

[56]    Supreme Court Act Reference, supra note 1 at para 74.

[57]    Constitution Act, 1982, supra note 3, s 41(d).

[58]    This may not have been an unintended feature of the Court’s opinion: it is in the Court’s institutional interest to read section 41(d) as broadly as possible and thereby immunize it from a host of changes by Parliament.

[59]    SC 1996, c 1, s 1.

[60]    See ibid, s 1.

[61]    It would not appear to apply to certain amendments under section 38 where provinces can opt out, and presumably not to those that only require the federal government acting alone. See Monahan & Shaw, supra note 39 at 217.

[62]    See ibid at 216–17.

[63]    See Howard Leeson, The Patriation Minutes (Edmonton: Centre for Constitutional Studies, 2011); Barry L Strayer, Canada’s Constitutional Revolution (Edmonton: University of Alberta Press, 2013); Russell, supra note 54 at 121, 141; “Premiers’ Conference, Ottawa, Ontario, April 16, 1981”, in Anne F Bayefsky, ed, Canada’s Constitution Act 1982 & Amendments: A Documentary History, vol 2 (Toronto: McGraw-Hill Ryerson, 1989) 804 at 804.

The Supreme Court in a Pluralistic World: Four Readings of a Reference in this issue Books Received

related content