Formal amendment rules have been the subject of great controversy in contemporary constitutional politics in Canada. From the patriation of the constitution in 1982, to the nearly ratified 1987 Meech Lake Accord and the similarly unsuccessful 1992 Charlottetown Accord, and through the Supreme Court’s Secession Reference in 1998, formal amendment rules have been at the centre of deep legal, political, and indeed moral disagreement in Canada.
The Supreme Court of Canada’s recent Senate Reference continues the modern trend. The constitutional questions on which the Court was asked to advise the Government of Canada focused squarely on the design and interpretation of the formal amendment rules entrenched in the Constitution Act, 1982. The questions concerned the requirements for imposing senatorial term limits, repealing the property qualification for senators, abolishing the Senate, and implementing a framework of consultative provincial elections that would inform prime ministerial nominations. In this article, I focus only on the last of these questions.
The nub of the issue concerned whether Parliament may constitutionally deploy its limited power of unilateral formal amendment under section 44 to make alterations to the method of prime ministerial nominations to the Senate of Canada, or whether Parliament is required to adhere to the more exacting multilateral formal amendment procedures defined in either sections 38 or 41. In my view, the answer was always clear. As I argued at the 2013 Constitutional Cases Conference at Osgoode Hall, before the Court rendered its advisory opinion, section 44—which authorizes Parliament to amend the Constitution of Canada “in relation to the executive government of Canada or the Senate and House of Commons”—is not the right vehicle for amendments to the method of filling vacancies in the Senate of Canada:
[T]he escalating and federalist structure of formal amendment entrenched in the architecture of Canada’s formal amendment rules suggests that it was illegitimate to use section 44 to make a formal amendment to an element of Canadian democracy as significant as senator selection. Using the default multilateral amendment rule in section 38 is more consistent with Canadian history, the evolution of the design of formal amendment rules in Canada, and the centrality of federalism to democratic self-government.
I suggested that the Government of Canada’s recourse to the unilateral formal amendment power under section 44 “reflects a disjuncture between legality and legitimacy.” I argued that although a purely formalist and strictly legalistic reading of the constitution could indicate that Parliament may pursue Senate reform through section 44, political history and constitutional design counsel that it would be illegitimate, whether legal or not. I concluded that
[i]n invoking this unilateral federal amendment power to formally amend senator selection, the Government of Canada has either misunderstood Parliament’s constitutional authority or attempted to achieve unilaterally what it is constitutionally required to pursue multilaterally.
The Court later concluded in its Senate Reference “that Parliament cannot unilaterally achieve most of the proposed changes to the Senate, which require [recourse to section 38].” Then as now, it was difficult to imagine the Court arriving at any other conclusion. The Court’s advisory opinion was constitutionally correct in its interpretation of the structure of formal amendment under the Constitution Act, 1982; it was well reasoned in its answers to each of the six reference questions; and it was politically prudent in requiring political actors to work cooperatively toward Senate reform pursuant to the text’s formally entrenched multilateral amendment procedures.
Yet in directing its attention methodically to the six reference questions, the Court missed an opportunity to bring to light the larger and more fundamental constitutional infirmities with the Government of Canada’s Senate reform ambitions.
I stress here that the Government of Canada’s Senate reform ambitions are not troubling in and of themselves. The Senate of Canada is in dire need of reform. As Ned Franks has recognized, the Senate is “a frustrating puzzle” and “the most criticized institution of government in Canada.” It is therefore with good reason that the Senate has been the subject of sustained debate since the adoption of Canada’s founding constitution in 1867. Only seven years into Confederation, the House of Commons was already debating Senate reform. Shortly thereafter, at the first intergovernmental conference, critics charged that the Senate was failing to meet the federalism-reinforcing objectives its designers had set for it. Paul Weiler subsequently captured the dominant sentiment of the twentieth century, observing that “just about everyone (except perhaps a few senators) would concede that the Canadian senate has not proved an effective representative of regional views in the central government.” The same critique endures today, as the Senate prepares to mark its 150th anniversary in 2017. Senate reform proposals themselves are therefore far from troubling—they are both welcome and necessary.
What is troubling, however, is how the Government has sought to pursue its Senate reform ambitions. The Government of Canada’s proposed Senate selection reforms concealed a broader strategy to innovate an unusual method of constitutional amendment: constitutional amendment by stealth. Constitutional amendment ordinarily channels public deliberation through formal, transparent, and predictable procedures designed to express the informed aggregated choices of political actors. But the Government of Canada appears to have calculated that the difficulty of formal amendment made its Senate reform objectives best achievable through informal and irregular procedures designed both to circumvent the textually prescribed rules for formal amendment and to introduce a material change to the Constitution of Canada. A constitutional amendment occurs by stealth when political actors consciously establish a new political practice whose repetition is intended to compel successors to conform their conduct to that practice. Over time, this practice matures into an unwritten constitutional convention that becomes informally entrenched in the constitution, though without the popular legitimacy we commonly associate with a constitutional amendment.
There are three distinguishing features of constitutional amendment by stealth—distinctions that make stealth amendment stand apart from other types of informal constitutional change: the circumvention of formal amendment rules, the intentional creation of a convention, and the twinned consequences of both promoting and weakening democracy. I explain each below in greater detail but a short word now may be useful. First, where political actors believe, correctly or not, that it is too difficult to use the formal amendment rules to entrench an amendment-level change, they resort to alternative informal means. Second, as a consequence of the difficulty of formal amendment, political actors circumvent the onerous formal amendment rules in the constitution’s text and opt instead to create a new democratic practice. Political actors intend this new practice to mature into a constitutional convention that will coerce their successors into compliance. Third, the convention that political actors seek to create is hard to resist because it enhances democracy. At its origin, though, the convention risks undermining democracy because it arises from a circumvention of the constitution’s formal amendment rules.
Here, the Government of Canada, which I will henceforth identify as the “Conservative government”, has long sought to replace senatorial appointments with democratic elections. But recognizing the virtual impossibility of formally amending the constitution to create an elected Senate, the Conservative government instead sought to create a political practice to achieve the same end: the prime minister would fill Senate vacancies only with candidates who had been endorsed in province-wide popular elections, pursuant to a parliamentary law creating a framework for consultative senatorial elections. As the democracy-promoting practice of voting for provincial senatorial nominees came to be viewed as encouraging civic participation and as enhancing the sociological legitimacy of the institution, future prime ministers would feel constrained to continue the practice, and that practice would over time mature into a constitutional convention. What would be lost in the entrenchment of this democratic convention is that it had been devised by political actors in response to the impossibility of creating an elected Senate in the only way the constitution permits: by formal amendment. This “stealth amendment” would have been simultaneously democracy-promoting and democracy-deficient: the politically expedient strategy to democratize the Senate would have given voters a voice in the selection of their senators, yet this democracy-promoting outcome would have been achieved by a constitutionally unsound circumvention of the textual rules for formal amendment.
As important as it is to make Canadian public institutions more democratic, the ends here would not have justified the means. An elected Senate is a worthy objective but not if the process by which we achieve it is itself devoid of democratic legitimacy. The formal amendment rules entrenched in the Constitution of Canada set the standard for legitimacy: in order to meet the test, political actors must follow the carefully detailed sequence and thresholds to make a material change to the constitution. The purposeful evasion of those rules undermines both the constitution and the new change created by the evasion. The Conservative government’s plan to establish, by unilateral amendment, a framework for provincial senatorial elections violated both the federalist principles underlying the constitution as well as the constitution’s peremptory rule that the method of senatorial selection may be amended only by multilateral formal amendment.
There were both intrinsic and instrumental reasons why the Conservative government sought to amend the constitution in this way. The constituents of the Conservative Party and its modern precursor parties, including the Progressive Conservative Party, Reform Party, and the Canadian Alliance, have long called for reforms to the Senate’s functions, seat distribution and method of appointment. The Conservative government’s recent proposals for Senate elections are an incremental step toward fulfilling its larger vision for institutional reform. As an instrumental matter, senatorial elections also give the Conservative Party some measure of insurance against the possibility of defeat in the House of Commons. Today, the Conservative Party enjoys the benefit of a fractured left, with the Liberal, New Democratic, and Green parties dividing the vote. But this vote split will not endure, just as the vote split on the right did not. In the event that parts or all of the left unite, the Conservative Party will see its institutional advantage evaporate, in which case it could lose the power to nominate senators. In the face of that contingency—a contingency that is attracting increasing coverage—the motivation for the Conservative government’s proposal for senatorial elections becomes clear: to decouple control of the House of Commons from control of the Senate, by requiring a direct vote for senators.
Whether these Senate reforms are a good idea is of little relevance to the fundamental question: are the means and ends of Senate reform consistent with the Constitution of Canada? Fortunately, the Court in the Senate Reference foiled the Conservative government’s bid to amend the constitution by stealth. But there is no guarantee that a future Court will rule similarly, nor that political actors will again seek permission or advice by way of a reference for answers to whether and how they may make small- or large-scale constitutional amendments. The failure of this stealth amendment is best viewed neither as a victory nor a loss for particular political actors but rather as an opportunity to learn about the pathologies of formal amendment under the Constitution of Canada, how those pathologies drive stealth amendment, and why Canadian political actors might find stealth amendment a profitable strategy for constitutional change. These are important questions that the Court did not address in its reference.
In this article, I fill the void in the Court’s careful yet incomplete advisory opinion by introducing, theorizing, and illustrating this unconventional form of constitutional amendment pursued by the Conservative government. I begin in Part I by examining constitutional amendment in Canada, with a focus on the difficulty of formal amendment, and the prevalence of informal amendment. In Part II, I explain the Conservative government’s Senate reform objectives with particular emphasis on the proposal to create a framework for senatorial elections. In Part III, I theorize constitutional amendment by stealth and explain how the Conservative government’s proposed senator selection reforms reflect an effort to amend the constitution by stealth, in circumvention of the deliberative procedures the constitutional text demands. I also draw from comparative perspectives to distinguish stealth amendment from other forms of informal constitutional change. I conclude Part III with attention to the costs and consequences of stealth amendment. In Part IV, I offer closing thoughts on constitutional amendment in Canada.
Formal amendment rules are fundamental to written constitutionalism. In constitutional democracies, formal amendment rules perform an essential corrective function: they authorize political actors to remedy discovered faults in the constitutional text in conformity with transparent procedures. At their best, formal amendment rules also distinguish constitutional from ordinary law, promote public discourse about constitutional interpretation, aggregate and translate popular preferences through public institutions, precommit political actors, check informal amendments that occur extra-constitutionally, and express constitutional values. Formal amendment rules therefore serve both symbolic and functional purposes.
In Canada, formal amendment rules serve each of these functions, at least in theory. But constitutional amendment in Canada is also special. Canada is one of only a few democratic constitutional states to entrench the restricted multi-track amendment framework, which assigns amendment procedures of varying difficulty to specific provisions or principles. It is equally worth noting that constitutional change in Canada has developed informally in unusual ways, not only as a consequence of evolving constitutional conventions, which is true of most if not all constitutional democracies, but more interestingly as a result of constitutional desuetude. Canada also finds itself among a shrinking number of countries without a theory or doctrine of unconstitutional constitutional amendment, something that only adds to Canada’s uniqueness.
What is most relevant for our purposes, however, is that the Constitution of Canada is one of the world’s most difficult constitutions to amend, earning the top score on Arend Lijphart’s index of constitutional rigidity. Yet what remains unappreciated about the difficulty of formal amendment in Canada is that the source of the constitution’s rigidity is not its formal amendment rules alone, which are admittedly exceedingly onerous. It is that those rules have been rendered even more demanding as a result of judicial interpretation, statutory enactment, and arguably also by constitutional convention. This extraordinary difficulty of formal amendment in Canada has consequently given rise to informal methods of amendment. In this Part, I explain and evaluate constitutional amendment in Canada with a view to setting the foundation for understanding how and why political actors would seek to innovate an unconventional method of informal amendment: constitutional amendment by stealth.
Measuring formal amendment difficulty is itself a difficult task. The limitations of existing cross-national formal amendment classifications illustrate the challenge of measuring amendment difficulty. As I have shown, some classifications are overinclusive, others are underinclusive, and still others are both, resulting in oversimplifications that elide important nuances that can either moderate or exacerbate formal amendment difficulty. For example, Edward Schneier’s important classification categorizes Canada with New Zealand and the United Kingdom, whose traditionally unwritten constitutions are associated with amendment ease, risking the false suggestion that Canada’s own formal amendment rules are similarly easy to satisfy. Donald Lutz has observed that Canada’s partially written and unwritten constitution poses a particular challenge for measuring amendment difficulty—namely how to determine what does or does not possess constitutional status in Canada—and this, in his view, makes it harder to measure amendment difficulty here than in most other constitutional democracies.
Formal amendment in Canada is “unusually complicated,” to quote an authority on amendment in the United States, which is thought to have one of the world’s most rigid constitutions. The strongest critique of the difficulty of formal amendment in Canada is the country’s lived experience: Canadian political actors succeeded in making historic constitutional changes in 1982, when Canada domesticated its constitution by creating made-in-Canada formal amendment rules, and entrenching the Canadian Charter of Rights and Freedoms. Yet history has shown that Michael Stein was right to question, shortly after the major multilateral constitutional changes in Canada in 1982, whether those changes would “prove to be only a Pyrrhic victory, a largely symbolic success that will effectively bring the process to a halt.” Since then, all major multilateral constitutional changes in Canada have met with failure.
Michael Lusztig’s theory of “mass input/legitimization” best explains why major multilateral constitutional amendment is virtually impossible today in Canada. Lusztig begins by observing that major amendment requires an extraordinarily deep and broad level of agreement by political actors. In addition to these demanding expectations, the prospect of major multilateral amendment efforts creates incentives for multiple constituencies to mobilize behind their interests in order to attain special status for themselves, and to entrench that status in the product of those amendment efforts. An additional complication results: the conferral of special status on one group makes it difficult to deny similar status to other groups. This leads to near-certain amendment failure for major multilateral formal amendment efforts involving fundamental or constitutive principles, the polity’s constitutional identity, or the framework and interrelations of public institutions. Lusztig and Christopher Manfredi therefore anticipate amendment failure for major multilateral amendments because political actors will make incompatible and intractable demands, both on the subject of the major amendment efforts themselves and on collateral issues of significance to their constituencies.
That Canada has five formal amendment procedures—each one more difficult to satisfy than the former—is more of a complicating than clarifying feature of its amendment rules. It is not always obvious which procedure must be used to formally amend a particular provision or principle, as the Senate Reference illustrates. The easiest procedure applies exclusively to formal amendments to provincial constitutions: the unilateral provincial procedure ensures that
the legislature of each province may exclusively make laws amending the constitution of the province.
Next, the unilateral federal procedure in section 44 authorizes the Parliament of Canada to unilaterally formally amend the constitution
in relation to the executive government of Canada or the Senate and House of Commons.
This procedure may not be used to amend matters expressly assigned to another, more difficult, amendment procedure.
The third amendment procedure—the parliamentary–provincial procedure in section 43—is deployable for regional matters. It must be used for formal amendments whose subject matter implicates “one or more, but not all, provinces”; for instance, an amendment concerning provincial boundaries, the use of English or French within a province, or the public funding of provincial religious schools. This threshold requires approval resolutions of both the House of Commons and the Senate, and of the unicameral provincial legislature or legislatures involved in the amendment.
The next most onerous procedure is the default multilateral amendment rule in section 38. It requires approval resolutions from both houses of Parliament, in addition to resolutions from the provincial legislatures of at least seven of Canada’s ten provinces, where the population of the ratifying provinces must amount to at least half of their total aggregate population. Political actors must use this threshold to amend senator selection and eligibility, Senate powers and provincial representation, the Supreme Court of Canada, proportional provincial representation in the House of Commons, and provincial–territorial boundaries. This threshold is the constitution’s default amendment rule, and political actors must therefore use it to formally amend all parts of the constitution not specifically assigned to another rule.
The most difficult formal amendment procedure is the unanimity threshold in section 41. It requires unanimous consent from federal and provincial political actors with approval resolutions from both the House of Commons and the Senate, and from each of the provincial legislatures. This threshold applies to the most important provisions and principles in Canadian constitutionalism: the structure and institutions of Canada’s constitutional monarchy; provincial representation in the House of Commons and the Senate, subject to related but lesser matters amendable by another specially designated lower threshold; the use of English or French, subject to the same qualification; the composition of the Supreme Court of Canada, subject again to the same qualification; and the entire structure of the formal amendment rules themselves. The architecture of formal amendment in Canada is therefore intricate in its escalating design.
Successfully adopting a multilateral formal amendment under either the general default or unanimity procedures requires constitutional politics to perform heroics. One scholar describes it as “largely impossible.” These demanding multilateral ratification thresholds have been satisfied only once since the entrenchment of Canada’s formal amendment rules over three decades ago. The overwhelming supermajority of amendments has occurred using the parliamentary–provincial and unilateral federal amendment rules, both of which are much easier to fulfill than the multilateral procedures designated for major constitutional changes. The most recent successful formal amendment occurred in 2011, when Parliament deployed its unilateral formal amendment power under section 44 to adjust the number of Members of Parliament consistent with the principle of proportionate provincial representation.
In Canada, as in other constitutional states, the difficulty of formal amendment has pushed constitutional change “off the books,” forcing political actors to update the constitution in ways that do not manifest themselves in a new constitutional writing. These unwritten changes, or informal amendments, alter the meaning of the Constitution of Canada in the absence of a textual modification. We can understand the impulse for informal amendment in terms of hydraulic pressure, as Heather Gerken describes it: where the path to formal amendment is blocked as a result of onerous procedures or unachievable majorities, the rigidity of the formal amendment rules will redirect the energies of political actors into alternative channels that will produce the same or similar outcome, albeit in a different form. These informal amendments may occur, for instance, by judicial interpretation, national legislation, executive decision, implication, convention, and desuetude. The functional result of an informal amendment is indistinguishable from a formal amendment insofar as both are binding on political actors.
In Canada, informal amendment has become the primary vehicle for constitutional change in the face of the near impossibility of formal amendment. As Allan Hutchinson explains, informal amendments “occur even though the formal process of constitutional change itself remains unused and unchanged.” The source for these informal changes has often been the judiciary. Courts, writes Hutchinson, “have become the preferred site for effecting important changes in the constitutional order.” Hutchinson remarks that informal amendment via judicial interpretation is a “less democratic” means for constitutional change than the legislative procedures authorized by the constitutional text, which require multiple expressions of popular will mediated by representative institutions: “in a society that claims to be devoted to the ideas and practice of democratic legitimacy, it is far from clear why the courts are the suitable or appropriate institution to speak and act on the people’s behalf.”
One need not agree with Hutchinson’s critique of the judiciary to recognize that informal amendment today prevails over formal amendment. Hutchinson demonstrates that
while almost none of the wording of the Constitution Act 1867 has changed in more than 125 years, the meaning and effect of its provisions on the division of provincial and federal powers have gone through a process of continuing redefinition.
For example, the historical interpretation of the “peace, order and good government power” shows that the meaning of a static constitutional text itself can change over time. Similarly, the interpretation of section 35 of the Constitution Act, 1982 in R. v. Sparrow prompts us to wonder “in what substantive, as opposed to formal, ways an amendment of the constitution would differ from it.” Many of these redefinitions have occurred by judicial interpretation, whose effect is virtually identical to an actual amendment. Indeed, writes Hutchinson, “it is difficult to pin down the difference in the substantive effects of the formal acts of amending the Constitution and the informal acts of interpreting it.”
There are two important qualifications to make to Hutchinson’s observations on the prevalence of informal amendment. Both concern the distinction between informal amendment and judicial interpretation. First, it is necessary to distinguish informal amendment by judicial interpretation from judicial interpretation itself, because not all judicial interpretation results in an informal amendment. The difference turns on the court in which the interpretation occurs: informal amendment by judicial interpretation occurs where the Supreme Court interprets the constitution as a final matter; in contrast, judicial interpretation by lower courts is generally not nationally binding and it is therefore less accurate to define it as an informal amendment. Second, even at the Supreme Court level, not all constitutional interpretation results in an informal amendment: the clearest case of an informal amendment by judicial interpretation occurs where the Supreme Court confers constitutional status upon an unwritten constitutional principle by subordinating duly passed legislation to that unwritten rule. In such a case, there is no functional difference in constitutional effect between a textual rule entrenched in the constitution by formal amendment, and an unwritten rule entrenched by judicial interpretation.
Canada’s already onerous formal amendment rules have themselves been informally amended to make them even harder than their text suggests. The Secession and Nadon References are two recent illustrations of the Court’s power to informally amend the Constitution of Canada, and more specifically, the constitution’s formal amendment rules. In both cases, the Court imposed additional constraints on political actors engaged in formally amending the Constitution of Canada. In the Secession Reference, the Court informally entrenched the duty to negotiate, and identified federalism, democracy, constitutionalism and the rule of law, and respect for minority rights as a handful of “underlying constitutional principles” that must govern the formal amendment process in connection with a provincial secession. This judicial interpretation amounts to an informal amendment insofar as these obligations now bind political actors in the same way they would were they formally entrenched in the constitutional text.
In the Nadon Reference, the Court exacerbated the difficulty of formal amendments to the Court itself. The Supreme Court informally entrenched its own essential features—which for the Court include, “at the very least, the Court’s jurisdiction as the final general court of appeal for Canada, including in matters of constitutional interpretation, and its independence”—against formal amendment by any other mechanism than section 41’s difficult unanimity procedure. Here, the Court’s interpretation clarified an open-textured constitutional provision that had once been susceptible to competing interpretations. Today, however, there is only one legal interpretation, and though it is not textually entrenched it is nonetheless binding on political actors.
Quite apart from their informal amendment by judicial interpretation, Canada’s formal amendment rules have also been informally amended by parliamentary, provincial, and territorial legislation. In 1996, Parliament adopted the regional veto law, which confers veto power to provinces or regions—the Atlantic provinces, Ontario, Quebec, the Prairie provinces, and British Columbia—in major constitutional reforms achieved via section 38. This veto law fulfilled the then-prime minister’s pledge to grant Quebec a veto on major constitutional reforms—a pledge made against the backdrop of Quebec’s near-successful referendum on secession in 1995. Under the law, no cabinet minister in the Government of Canada may propose a constitutional amendment pursuant to the multilateral amendment procedure in section 38 without first securing the consent of a majority of the provinces, including British Columbia, Ontario, and Quebec, along with Alberta by implication of the current distribution of provincial population. Although the regional veto law is an ordinary statute that may be repealed by an ordinary law, it nevertheless constrains the formal amendment process by adding a requirement that is not written into the master text of the Constitution of Canada, but is now effectively informally entrenched within it. Relatedly, provinces and territories have adopted their own laws on national formal amendment: several now require a binding or advisory referendum on a multilateral formal amendment proposed by Parliament before holding a ratifying vote in their provincial or territorial legislature.
In addition to the onerous escalating formal amendment rules, as well as the constitutionally uncodified judicial and legislative requirements layered onto them, formal amendment in Canada may also be further complicated by constitutional convention. It has been suggested that the existing amendment thresholds do not reflect the new expectations of popular participation in constitutional amendment. The decision to require a national referendum to ratify the failed Charlottetown Accord may have created an expectation that future large-scale amendment efforts will also require one, making formal amendment even more difficult than its textual thresholds suggest. Another convention has arguably emerged from the same Charlottetown referendum: the territories’ involvement in the 1992 referendal process may have created a precedent requiring their future participation. For all of these reasons, multilateral formal amendment on major constitutional issues in Canada is difficult, to say the least.
The difficulty of formal amendment has made it unlikely to achieve Senate reform using the multilateral amendment procedures entrenched in the Constitution of Canada. It is also improbable to achieve meaningful Senate reform through the informal amendment process of judicial interpretation: the Court cannot, by interpretation, order elections for senator selection, nor can it impose a more equitable distribution of Senate seats, nor mandate more democratic measures in the Senate’s internal operation to make it more effective. In this way, the structure and design of the Senate are what Sanford Levinson would describe as “hard-wired” features of the constitution that cannot be changed though the process of informal amendment.
Before the adoption of the Constitution Act, 1982, Canadian political actors updated the constitution in the absence of formal amendment rules under a constitutional convention requiring provincial consent for significant constitutional changes. The Constitution Act, 1982 finally entrenched formal amendment rules that more clearly authorized formal changes to the Senate, but political actors failed on two occasions to reform the Senate within larger frameworks of constitutional revision: first in the 1987 Meech Lake Accord, and next in the 1992 Charlottetown Accord. These two momentous failures of large-scale constitutional revision did not quell calls for Senate reform, certainly not from western Canada, the origin of proposals for a Triple-E Senate. Indeed, most Canadians have long supported, and continue today to support, some form of Senate reform. But in light of the recent failures of wholesale constitutional renewal, it is difficult to see a path to Senate reform by formal amendment.
The Conservative government’s modern proposals for Senate reform—the predicate for the Supreme Court’s Senate Reference—provide a contrast to the earlier, more comprehensive formal efforts to amend the Senate. As Bruce Hicks writes, whereas “substantive Senate reform has failed to get traction since Confederation,” the current approach toward Senate reform suggests a preference for “incremental changes that either do not require amending the Constitution Act, 1867 or do not require provincial consent.” In the Part to follow, I will argue that the modern proposals are just as significant as earlier efforts to reform the Senate, only that the modern proposals—had the proposals been validated by the Supreme Court, and subsequently entrenched into law—would have amended the Constitution of Canada informally, though with similarly far-reaching effect as the formal amendments envisioned by earlier large-scale attempts. First, however, it is important to understand precisely what the modern reforms proposed to do.
The questions referred by the Conservative government ask the Supreme Court to evaluate Parliament’s power to pass three Senate reform bills. The Senate Term Limits bill, introduced in 2006, proposes to formally amend section 29 of the Constitution Act, 1867 by establishing an eight-year term limit for new senators. The Senate Appointment Consultations bill, introduced in 2007, proposes to create consultative provincial and territorial elections to gauge voter preferences to fill Senate vacancies as they arise in their province or territory. The Senate Reform bill, itself introduced in 2011, combines parts of the Senate Term Limits and Senate Appointment Consultations bills into one: it establishes a nine-year term limit for new senators by formally amending section 29 of the Constitution Act, 1867, and it moreover creates a framework for provincial and territorial elections to fill Senate vacancies. In this article, I focus only on the framework of advisory elections that the Conservative government proposed to create without a formal amendment. It is my claim that this framework for consultative senatorial elections would have informally amended the Constitution of Canada had the Court authorized Parliament to create this framework.
The Senate Reform and the Senate Appointment Consultations bills create a similar framework for consultative senatorial elections. The purpose of both bills, stated in the former though not in the latter, is to constrain the prime minister to appoint senators who claim the popular support of voters in their province or territory. The Senate Reform bill’s primary governing principle holds that “[s]enators to be appointed for a province or territory should be chosen from a list of Senate nominees submitted by the government of the province or territory,” with the list of nominees “to be determined by an election held in the province or territory.” The bill stipulates that where a province or territory has adopted the proposed electoral framework, which is set out in Schedule 1, the prime minister “must consider names from the most current list of Senate nominees selected for that province or territory” when “recommending Senate nominees to the Governor General.” The electoral framework contains instructions for administering consultative senatorial elections in a province or territory, including rules about candidate eligibility, election timing, election administration officials, nomination procedures, balloting, concurrent provincial, territorial, or municipal elections, as well as other regulations attendant to holding elections. The Senate Reform bill also establishes a nine-year term for senatorial appointees. Whereas this new term limit is identified as an express amendment to section 29 of the Constitution Act, 1867, the new framework for consultative senatorial elections is not expressly identified as an amendment to the Constitution of Canada. This is a problematic omission, as I will show in Part III.
The Senate Appointment Consultations bill has the same objective: to constrain the prime minister to consider provincial or territorial consultative senatorial election winners for appointment to the Senate. The bill’s framework for consultative elections adopts many of the same elements as the Senate Reform bill. The Senate Appointment Consultations bill states generally that the Government of Canada is “committed to pursuing comprehensive Senate reform to make the Senate an effective, independent, and democratically elected body that equitably represents all regions.” Yet the bill appears to concede its own legal precariousness when it states that it seeks to alter the way the prime minister makes recommendations to the Governor General for eventual appointment “pending the pursuit of a constitutional amendment under subsection 38(1) of the Constitution Act, 1982 to provide for a means of direct election.” Specifically, the bill creates “a method for ascertaining the preferences of electors in a province on appointments to the Senate within the existing process of summoning senators.” Although there may be a formal difference between establishing direct senatorial elections via section 38, and creating consultative senatorial elections as the bill intends to establish, there is little functional difference between the two types of selection mechanisms. The bill expressly identifies the long list of statutory amendments that would result from these major changes to senatorial selection but says nothing of the resulting constitutional amendments. As I will show in Part III, this is a problem for the democratic legitimacy of the consultative senatorial elections the Conservative government sought to establish.
In the recent Senate Reference, the relevant question for the Supreme Court was “whether Parliament, acting alone, can reform the Senate by creating consultative elections to select senatorial nominees endorsed by the populations of the various provinces and territories.” The Court rejected both major arguments from the Conservative government, namely, first, that the introduction of consultative elections does not “constitute an amendment to the Constitution of Canada in relation to the method of selecting senators,” and “in the alternative that, if the implementation of consultative elections requires a constitutional amendment, then it can be achieved under the unilateral federal amending procedure (s. 44).” The Court concluded that the Conservative government could not proceed as planned to establish its framework for provincial consultative elections to fill senatorial vacancies without violating the Constitution of Canada. It is important to understand the Conservative government’s two major arguments.
The Conservative government’s first major argument was that creating a framework for consultative senatorial elections does not constitute an amendment to the Constitution of Canada. It does not amend the constitution because consultative elections may be introduced without changing the text of the constitution, where the formal process for appointing individuals to the Senate—specifically by official summoning by the Governor General on the advice of the prime minister, as required by the Constitution Act, 1867—remains unchanged.
The Court advised the Conservative government that it could not accept this argument because it “privileges form over substance.” A constitutional amendment is more than a formal amendment to the constitutional text, and were the Court to accept the first major argument, it would “[reduce] the notion of constitutional amendment to a matter of whether or not the letter of the constitutional text is modified.” Although introducing consultative elections would not change the constitutional text, “the Senate’s fundamental nature and role as a complementary legislative body of sober second thought would be significantly altered.” The Court made two basic points in this respect. First, relying on consultative elections to fill senatorial vacancies “would amend the Constitution of Canada by fundamentally altering its architecture.” The Senate would be transformed from an appointed body, designed to bring moderation and deliberation to the legislative process, to an elected body that would lose its independence from the electoral process. Such a transformation would introduce partisanship into the Senate’s legislative role, and thereby undermine its moderating and deliberative functions, and risk making the Senate an adversarial rather than complementary chamber to the House of Commons.
The Court also gave a textual reason why creating consultative elections would constitute a constitutional amendment. For the Court, the escalating structure of formal amendment leaves little room to doubt that the multilateral amendment procedure in section 38 must be used to make constitutional amendments in connection with “the method of selecting Senators,” to quote directly from the constitutional text. This generalist language on constitutional changes to senatorial selection “covers the implementation of consultative elections, indicating that a constitutional amendment is required and making that amendment subject to the general procedure [in Section 38].” The Court again invoked the distinction between form and substance to support its interpretation of the constitution: “[t]he words ‘the method of selecting Senators’ include more than the formal appointment of Senators by the Governor General.” The constitution’s drafters chose this language in order to cover all alterations to the method of senatorial selection, not only for changes to the means of appointment. Therefore, explained the Court, the new framework of consultative elections, which would create candidate lists from which the prime minister would be expected to choose nominees for senatorial appointments, would effectively change the method of senatorial selection, and would therefore constitute an amendment requiring recourse to the multilateral amendment procedure in section 38.
The Conservative government’s alternative argument conceded that creating consultative elections would constitute an amendment to the Constitution of Canada but insisted that such an amendment could be achieved using the unilateral federal amendment power under section 44. Subject to sections 41 and 42 of the Constitution Act, 1982, Parliament is authorized under section 44 to “exclusively make laws amending the Constitution of Canada in relation to the executive government of Canada or the Senate and House of Commons.” The Conservative government therefore argued that introducing a framework of consultative elections would constitute an amendment to the Constitution of Canada “in relation to … the Senate” that is achievable through Parliament’s amendment authority in section 44.
The Court likewise rejected this argument using a textualist interpretation of the constitution. Since Parliament’s unilateral amending power is expressly made subject to section 42—which applies the multilateral amendment procedure under section 38 to changes to “the method of selecting senators”—it follows for the Court that section 38 must be used to initiate consultative elections. Section 42 is peremptory in removing its designated items—for example, amendments to the Supreme Court of Canada, the principle of proportional representation in constituting the House of Commons, and importantly to senatorial selection—from the parliamentary power under section 44. To read the interrelation of sections 38, 42, and 44 otherwise would be to misread the limited scope of section 44, and to confer upon Parliament the unilateral power to fundamentally change the Constitution of Canada without the required measure of provincial consent contemplated by the escalating structure of formal amendment. Therefore, wrote the Court, Parliament cannot lawfully introduce consultative elections, and thereby amend the Constitution of Canada using the narrow power of section 44.
It is difficult to find error in the Supreme Court’s careful rejection of the Conservative government’s proposed Senate reforms. As an exposition of legal doctrine, it is consistent with the standard set by the Court’s best precedents in defining constitutional law. As an act of judicial statecraft, it achieves two important objectives: it sets the rules for Senate reform in Canadian constitutional politics, and confirms the Court as the ultimate arbiter of constitutional meaning. But as an exercise in constitutional statesmanship, the strongest critique of the Court’s advisory opinion is precisely that it is carefully measured, perhaps too much so. The Court failed to expose the extent to which the proposed introduction of consultative senatorial elections—not the consultative elections themselves, but the way the Conservative government proposed to introduce them into Canadian political practice—violates not only the constitution, but also the democratic values of transparency, accountability, and predictability in the rule of law.
The Court made no mention of the democratic deficiencies in the Conservative government’s plans to institute consultative senatorial elections. It did, however, suggest that consultative elections might create a constitutional convention on senatorial selection and nomination. The Court suggested that requiring the prime minister to consider the individuals identified on the lists of elected candidates for appointment to the Senate would tie the prime minister’s hands in senatorial selection. Acknowledging that “[i]t is true that, in theory, prime ministers could ignore the election results and rarely, or indeed never, recommend to the Governor General the winners of the consultative elections,” the Court reasoned correctly that “[w]e cannot assume that future prime ministers will defeat [the purpose of giving senators a popular mandate] by ignoring the results of costly and hard-fought consultative elections.” But the Court could have said more to lay bare the problematic irregularity of the procedures by which the Conservative government sought to introduce consultative senatorial elections.
Others have suggested similar critiques of the proposal for consultative senatorial elections. For example, the Canadian Bar Association has opposed consultative elections because although they would “not affect the legal authority of the Prime Minister to select nominees to be appointed to the Senate,” they would “affect his or her practical ability to select such nominees.” The Association explains that the prime minister’s practical authority would be constrained as he or she began to draw nominees from lists of elected candidates because it would become difficult “to choose any candidate other than those preferred by the consultation vote.” The prime minister would understandably “be reluctant to ignore the direct expression of the electors,” and thus “[i]t is conceivable that the conditions for the creation of a convention may arise over time such that a Prime Minister would always respect the electorate’s choice.”
Scholars have made related, though different, critiques. Fabien Gélinas and Léonid Sirota, for instance, have argued that consultative senatorial elections would increase the political authority of the Senate contrary to its original design. This could eventually overrun the convention limiting the Senate’s power to oppose the will of the House of Commons, with consequences for responsible government, Cabinet formation, Canada’s constitutional monarchy, and, more generally, the Constitution of Canada. José Woehrling has suggested that it would be preferable to reform senatorial selection through the establishment of a constitutional convention pursuant to which the federal executive would select senatorial nominees on the recommendation of provincial executives. For his part, Gary O’Brien, the former Deputy Clerk of the Senate, has cautioned that consultative elections could upset current committee operations in the Senate, and would in turn change its nature and function.
These are strong criticisms of the Conservative government’s Senate reform proposals, but none addresses squarely the core of the democratic deficiency in consultative senatorial elections. Consultative elections are not problematic in and of themselves. On the contrary, they reflect a potentially positive step toward demystifying the Senate. However, their proposed introduction through irregular means is democratically deficient in important ways that threaten to weaken the rule of law in Canada. This is perhaps a paradoxical argument—after all, how could an indirectly elected Senate weaken the rule of law or reveal democratic deficiencies?—but it is worth making in defense of transparency, accountability, and predictability. The process by which the Conservative government proposed to introduce consultative senatorial elections would have undermined democratic values, notwithstanding the democracy-enhancing result of consultative elections replacing the existing practice of non-democratic senatorial appointment with a more democratic method of electoral consultation. In the Part to follow, I turn my attention to the democratic deficiency in the Conservative government’s proposal for consultative senatorial elections.
Constitutional amendment ordinarily channels public deliberation through formal, transparent, and predictable procedures designed to express the informed aggregated choices of political actors. The Conservative government’s plan to create consultative senatorial elections amounted to a proposal for a major constitutional amendment to the basic structure of the Constitution of Canada, yet it was designed to occur outside of this conventional process of constitutional amendment. The Conservative government deployed a strategy of stealth amendment through informal and irregular procedures in order to circumvent the onerous (but nonetheless constitutionally required) rules for amending senatorial selection. Constitutional amendment by stealth is a species of informal amendment. It has three distinguishing features: first, it is an effort to circumvent the rules of formal amendment; second, its underlying intent is to create a practice that will ultimately bind successors to a constitutional convention; and third, the substance of the convention itself is democracy-promoting but its origins are democracy-deficient insofar as they circumvent the constitution’s formal amendment rules.
In this Part, I explain the concept of constitutional amendment by stealth, and illustrate how the Conservative government’s proposed senator selection reforms reflect this method of informal and irregular constitutional change. I also explore the costs and consequences of stealth amendment. My ultimate claim is that the stealth senatorial selection amendment attempted by the Conservative government violates the democratic rule of law values of transparency, predictability, and accountability. Let me stress, at the outset, that much of this analysis is a modest effort to project into the future what would have happened had the Conservative government passed its framework for consultative senatorial elections into law. The point is to explore the implications for constitutional law and politics of consultative senatorial elections.
The best critique of the Conservative government’s Senate reform efforts concerns what Mark Walters calls the strategy “to exploit the distinction between constitutional law and constitutional convention, and to legislate for an elected Senate within the realm of convention while leaving the appointed Senate in tact as a matter of law.” Walters suggests that the consultative elections law, if passed and sanctioned by the Court, would have ultimately compelled the provinces to opt in to the system of provincial consultative senatorial elections. The source of compulsion, however, would not necessarily be a convention, explains Walters, but would rather stem from political necessity.
1. The Constraint of Consultative Elections
To understand Walters’s sophisticated argument, it is important to accept his point of departure, namely that there is a distinction between feeling “obliged” to do something and feeling “obligated” to do that thing. The creation of a constitutional convention, writes Walters, turns on the “special sense of obligation” that binds political actors to feel “obligated to do” something, a feeling that is “necessary for there to be a rule,” without which “no true constitutional convention could be said to exist.” In absence of such an obligation, the consultative elections law would not create a proper rule but, rather, simply a practice. Walters even doubts whether “a true convention on appointing elected senators could emerge” in the future. He concludes that “[t]he real point of the legislation, then, is to make it politically difficult, perhaps even impossible, for future prime ministers to depart from this practice.” It would become politically unpalatable for the prime minister to refuse to nominate the victorious candidates, because the duty to consider the consultative senatorial elections would leave the prime minister with no choice but to nominate the candidate chosen by voters. Walters illustrates the unavoidable constraint that the political reality of consultative elections would bring to bear:
Confronted by the statutory duty to “consider” election results, a future Prime Minister who disagreed with, and wanted to depart from, Harper’s practice of appointing elected senators to an otherwise unreformed Senate would either have to “consider” the election results and reject them, appointing someone the people did not select in place of the people they did select, or take steps to have the legislation repealed, thus appearing to take away from the people a right to vote for their legislators, and it is fair to assume that selecting either course of action would involve political costs that the departure from a mere practice unsupported by legislation would not.
This is problematic, writes Walters, because consultative elections appear to “[establish] the basis for a constitutional trap not a constitutional convention.” In the short term, the Senate would in practice, though not by convention, become elected, but it would remain unreformed both in terms of its seat distribution and its constitutional powers and functions. In the long term, however, the Conservative government would seek, first, to formalize this practice with a formal amendment to constitutionalize an elected Senate, which by then will have become an unchangeable political reality; and second, to force the provinces to negotiate the other aspects of Senate reform with constrained choices on reform possibilities since an elected Senate will have become a “foregone conclusion.” Walters stresses this idea of compulsion, or force as he emphasizes, in lamenting that the Conservative government’s introduction of consultative elections would have rigged the rules of future constitutional change:
Once the step toward an elected Senate is taken, the basic course of future reform will have been established and the choices available to actors involved in subsequent steps will have been forced in ways that they would not otherwise be.
Walters is justifiably concerned about consultative elections. But even Walters’s strong critique of consultative elections is incomplete. Walters is correct that consultative elections would set a “trap” for the provinces in future negotiations on constitutional reform, but it is important to identify the actors, objects, and subjects of the entrapment more clearly than he does. Second, I depart from Walters’s argument on whether the introduction of consultative elections as proposed by the Conservative government would ultimately create a constitutional convention binding upon future prime ministers. Walters argues no, but there is a much stronger case to be made than Walters suggests. Third, as I will discuss in the section to follow, had the Supreme Court approved the Conservative government’s proposal for consultative elections, the outcome would have been harmful to democracy and the rule of law in Canada—a paradoxical point, to be sure, given that consultative elections would have given Canadians the power to vote for their heretofore appointed senators.
2. The Actors, Objects, and Subjects of Entrapment
Walters is not specific enough about the effect of consultative elections on the provinces, nor does he explain how much further than the provinces themselves the effect would extend. First, it is undeniable that consultative elections would limit the range of choices available to provincial premiers, provincial legislatures, and also of the provincial electorate in future constitutional reform. The prime ministerial practice of nominating winning consultative election candidates would become politically irreversible despite possible opposition from provincial premiers, who would risk losing their status as the province’s voice in Ottawa in favour of elected senators sent by voters to Ottawa for that purpose. Provincial legislatures, for their part, would have a difficult time justifying any opposition to proposals to formalize an elected Senate given that consultative elections would have effectively led to a de facto elected Senate. And, for the same reason, the provincial electorate would not accept anything less than an elected Senate in formal constitutional negotiations. The negotiations on any future formal amendment would therefore be distorted by the informal transformation of the Senate into a de facto elected body.
But the effect of consultative elections would extend beyond the provinces themselves. The prime ministerial practice would bind future prime ministers to follow his example, which as I will argue below would eventually mature into a convention. It would deny other political actors, including the parliamentary opposition, provincial premiers and legislatures, as well as the Canadian electorate, the constitutional right to deliberate on whether Canada should have an elected Senate—a right that is in any case virtually meaningless without the concurrent authority simultaneously to make substantive changes to the powers and functions of the Senate. The subjects of the Conservative government’s entrapment would therefore be the entire universe of Canadian political actors. When the time arrived to constitutionalize senatorial elections and to make related Senate reforms, Canada’s federalist safeguards to major constitutional change would be just one of many constitutional rules obviated by the Conservative government’s by-then normalized practice of consultative senatorial elections.
There is an additional point worth making: the purpose of the Conservative government’s entrapment was unconstitutional at best, and illegitimate at worst. The Conservative government sought to do unilaterally what it could not do multilaterally. As I will explain in detail below, any future constitutional change to the Senate, and to the Constitution of Canada, would have followed from the framework of consultative elections adopted by Parliament alone with no further national consultation, and it would have been operationalized exclusively by the prime minister in his choice of whom to nominate to the Senate. One person would therefore have had a disproportionate influence on the reform of the Senate, contrary both to the actual design of the constitution, which mandates multilateral agreement for a constitutional change of such significance, and to the spirit of the constitution, whose architecture is intended to foster cooperative federalism, not executive constitutionalism, in major constitutional change.
What made the plan for senatorial elections devious is what made it brilliant: the Conservative government sought to “trap,” to capitalize on the term used by Walters, provincial actors into no other alternative but to ratify by future constitutional amendment the framework for an elected Senate created by consultative elections—whether or not the provinces indeed supported the idea on its merit. The trap would have been inescapable: either formally entrench consultative elections or deny the provincial electorate the de facto right to elect its own senators, a right that voters would have deemed vested in light of their continuing practice of electing their senators. But there is a stronger case than Walters suggests that consultative elections as proposed by the Conservative government would have created a constitutional convention binding future prime ministers, though binding only politically, not legally.
3. The Convention on Senator Selection
A convention, which is an obligation to act “in a way other than what the formal law prescribes or allows,” can arise from practice, agreement, declaration, or principle. Had the Court approved consultative elections, the origin of the convention requiring the prime minister to nominate the consultative election winners could not have been traced to either practice or principle alone. Moreover, the lack of public agreement ex ante or ex post to the convention, as well as the absence of any authoritative declaration that a convention was being established, would have been problematic, as I explain below. The meaning of a convention and its formation are key to understanding why consultative elections as proposed by the Conservative government would have ultimately created a convention requiring future prime ministers to conform their conduct to the precedent of nominating the consultative election winner to the Senate.
The study of conventions must begin with Ivor Jennings’s three-part test for their creation. Jennings wrote that “[w]e have to ask ourselves three questions: first, what are the precedents; secondly, did the actors in the precedents believe that they were bound by a rule; and thirdly, is there a reason for the rule?” This test requires precedents for establishing a convention, political actors to feel bound by those precedents, and a normative reason for the rule supported by conventional practice. As to the first question, Jennings explained that “mere practice is insufficient.” As to the second and third, just because political actors do act a certain way does not mean that they should; they must “believe that they ought to do so” in order for a convention to exist. The creation of a convention must also “be due to the reason of the thing because it accords with the prevailing political philosophy,” meaning that it “helps to make the democratic system operate,” or that “it enables the machinery of State to run more smoothly.” And where the convention continues to operate “because it is desirable in the circumstances of the constitution, it must be created for the same reason.” These then are the three conditions for the creation of a convention: precedent, self-consciousness, and normativity.
A convention may arise separately in four ways. It may arise as a result of a political practice seen as “necessary to protect some facet of the constitutional system,” by agreement where “the main political actors” expressly create or alter a “sort of contractual agreement” to act in a certain way, from the intent to establish one “by authoritative unilateral declarations by key political actors,” and on principle where “a substantive obligation” exists “requiring political actors to behave in a certain way.” In light of the historically harsh critique of the Senate for its democratic deficit as an appointed chamber, the Senate would have accrued a new democratic legitimacy as an elected body under the plan for consultative elections, and this would have been difficult for prime ministers to ignore. Refusing to nominate a senatorial election winner would have invited the disapproval of voters who would have seen the previous prime minister accede to their democratically expressed wishes to choose their senators. The continued prime ministerial nomination of election winners would have become a de facto moral obligation to respect the democratization of the Senate set in motion by the elections themselves.
Assume the Court had reached the opposite conclusion in the Senate Reference: that the Conservative government has the constitutional authority to create a framework for senatorial elections using section 44. Under the majority Conservative government, the bill would have passed in both houses, it would have received royal assent, and it would have come into force relatively soon—let us say no later than the end of 2014. With twenty Senate vacancies across seven provinces, it is not unreasonable to posit that there would have been pressure on both federal and provincial political actors to fill them. At least some, and perhaps all, of these provincial consultative senatorial elections would have been held prior to the federal general election, scheduled for October 2015, and the current prime minister, exercising his personal prerogative to choose whom to appoint, would have nominated the winning candidates who would in turn have been summoned to the chamber by the Governor General, as the constitution requires.
Whether or not the incumbent prime minister had won re-election in the general election, this practice of prime ministerial nomination of consultative election winners is likely to have continued. It would have either persisted under the re-elected prime minister, or under a new prime minister from the incumbent party, or, perhaps grudgingly, under a new prime minister from the previous opposition. Under the law authorizing provinces to hold elections to choose their Senate representatives, those elections would have yielded clear indications of voter preferences for Senate nominees in their province. The current prime minister, if re-elected, would have continued the practice, as it had been his declared preference. A different prime minister would only at his or her peril have cast aside the clearly expressed wishes of voters, even if the province had not been one that tended to support his or her party. Failure to heed the choice of provincial voters would have given the opposition ammunition to deride the prime minister, and it would have moreover caused members of his or her own party to question the commitment to reforming the Senate, and more broadly, to democratizing public institutions in need of change. It is, of course, possible that a new prime minister would have resisted the practice begun by his predecessor to nominate winning candidates to the Senate. But the new prime minister would still have had to contend with provincial leaders and voters who would have come to expect, from earlier nominations, the right to continue choosing their senators.
The expectation of the right of choice created by prior practice would moot inquiries into whether the right had been properly created. The regularity of the practice would cause politics to override law, transforming a practice into a conventional right over time as political actors continued to engage in it. Political pressure to conform to prior practice would change the rule of recognition to recognize the validity of the expectation that prime ministers will nominate the winning consultative election candidates to the Senate. In Hartian terms, prime ministerial practice of nominating the consultative election winner to the Senate would become “a rule of the group to be supported by the social pressure it exerts.” That a convention arises in this way does not undermine the force or legitimacy of the convention, as long as political actors self-consciously act in a way reflecting their “shared acceptance” of the practice as a “guiding rule” for their conduct. Conventions, after all, “ultimately reflect what people do,” and they are the result of political actors internalizing a rule as obligatory.
What makes it even harder to imagine that the governing party could discontinue this practice of prime ministerial nomination of consultative senatorial election winners is that the practice is supported by the principle of democracy. This speaks to the normativity that Jennings insists must underpin a political practice before it becomes a convention. Although the process by which the convention had been established would have belied the formal rules for changing the method of senatorial selection, it would have become cloaked in a nearly unassailable democratic legitimacy that can be conferred only by free electoral choice. The newly democratized Senate would have become untouchable: “an area in which the freedom of the actors on the governmental stage is curtailed (though not by legal restraints)” and in which those actors “are precluded from adopting the policy that accords with their perception of what the public interest requires.” Here, the democratic principle justifying senatorial elections would have frustrated any inclination that opponents might have had to deny voters the acquired right to choose their senators. It is in this way that the practice would have become a convention: after a critical number of exercises, “[a]ny deviation from the practice attracts—and is rightly regarded as attracting—criticism and pressure to conform.” There would henceforth have been no reasonable political basis for abolishing senatorial elections, although the legal case would be strong in light of its constitutionally illegitimate origins, as I discuss further below.
It is important to stress that the practice would not have matured into a constitutional convention without the compliance of opposition parties. Cross-party ratification, either by affirmative approval or grudging acquiescence, is a condition of the creation of a convention. The real essence of a convention, Joseph Jaconelli explains, “is to be found in the system of concordant actions and expectations that draws into its compass even those who were not parties to the agreement.” The test for Jennings’s second of the three questions—whether the actors in the precedents believe they were bound by a rule—can be answered definitively only where the opposition, when it attains power, conforms its conduct to a practice established by its predecessor. As I have suggested above, it would be politically unpalatable for opponents to discontinue the practice of prime ministerial nomination of winning consultative election candidates. The self-conscious ratification of the practice by the former opposition would validate the practice by mutuality of approval—a practice whose continued observance across parties would ultimately transform it into a convention legitimated by cross-party precedent.
Though it would have arisen by practice and principle, the convention of prime ministerial nomination of winning consultative senatorial election candidates would have lacked agreement and declarative transparency. It would have lacked the former because we know that the Conservative government’s proposed framework of consultative elections is opposed by the opposition and across many provinces. It would moreover have lacked declarative transparency because the prime minister did not speak at the time of its tabling in the House, nor has he since spoken, of the introduction of consultative senatorial elections as a way to create a convention that will bind his successors in whom to nominate to the Senate. On the contrary, the Conservative government argued before the Supreme Court in the Senate Reference that future prime ministers would retain their discretion to choose whom to nominate to the Senate and would not be compelled to nominate the consultative election winner. The absence of agreement and declarative transparency highlights the irregularity of this convention, though it does not in the end undermine its binding quality upon political actors. That the convention would be binding as a political matter but lacking in legal basis raises a question in need of a different answer: how, precisely, would this new convention be constitutionally deficient?
It is reasonable for Walters to predict that the introduction of consultative elections could in the long term rig the rules of constitutional change so as to limit the range of reform options available to political actors. After years of unbroken precedents of prime ministers nominating for senatorial appointment the winners in province-level consultative elections, future prime ministers would become expected to honour the practice, which, as I have argued above in contrast to Walters, will have matured into a constitutional convention. The origin of this convention would be traced to political necessity but the convention itself, which Walters regards as a practice, would in short course reflect an intrinsic democratic value.
The consequence of introducing consultative elections is even more problematic than Walters perceives. Consultative elections could admittedly tie the hands of political actors in future constitutional reforms, but that is only one of their adverse consequences. The more important one is that the consultative elections proposed by the Conservative government would have undermined the rule of law. Consultative elections themselves are not the problem; it is rather the way the Conservative government sought to introduce them. By pursuing its reform efforts in defiance of the textually required formal amendment rules for altering the method of senatorial selection, the Conservative government attempted to circumvent the constitution’s public, transparent, and predictable procedures for making changes to its basic structure. Had the Conservative government succeeded in creating its new framework for consultative elections, it would have been a constitutional amendment by stealth.
Constitutional amendment by stealth is an innovative but illegitimate method of constitutional change. It occurs where political actors consciously establish a new political practice whose repetition is intended to create an expectation that successors will have to comply with that practice as it matures into a constitutional convention. Constitutional amendment by stealth is driven by the political reality that formally amending the constitution is difficult if not improbable. In light of the near or actual impossibility of formal amendment, political actors choose to ignore the formal rules of the constitution and instead pursue their reform objectives through informal and irregular procedures, all with the intent of submitting their successors into compliance. Constitutional amendment by stealth, therefore, deliberately evades the public, transparent, and predictable formal amendment procedures that are designed precisely to express the informed aggregated choices of multiple political actors rather than the preferences of a few.
1. The Values of the Rule of Law
The Supreme Court of Canada’s conception of the rule of law gives primacy to the constitutional text. As the Court has held, the rule of law “requires that courts give effect to the Constitution’s text, and apply, by whatever its terms, legislation that conforms to that text.” This positivist interpretation of the rule of law is not entirely procedural. The Court has recognized that the rule of law embraces three principles: the supremacy of law over both public and private actors, the legal regulation of interactions between public and private actors, and the establishment and maintenance of positive laws that reflect an order of normative values. Importantly, however, the Court has stressed that although law must be sustained by normative values, any discovered principles should cohere with the text, not undermine it, because “[t]he rule of law is not an invitation to trivialize or supplant the Constitution’s written terms.”
Constitutional amendment by stealth violates the democratic values of the rule of law. It does not satisfy the rule of law’s expectations of transparency, accountability, and predictability. In the most influential scholarly articulations of the rule of law, Lon Fuller, A.V. Dicey, and Friedrich Hayek each separately stress the discretion-limiting quality of the rule of law and its cornerstone feature of consistency between the law as written and as applied. For Fuller, the rule of law requires a legal system to respect at minimum eight criteria, four of which appear to be infringed in a material way by stealth amendment. First, the rule of law rejects lawmaking created “on an ad hoc basis.” Law must instead spring from formal procedures allowing opportunities for open and meaningful deliberation about its implications. Second, the rule of law rejects the “failure to publicize” the laws relied upon by political actors. When the law is not known, it cannot be properly followed, understood, or challenged as to its constitutionality. Third, the rule of law rejects the “failure of congruence between the rules as announced and their actual administration.” The rules on the books should as much as possible match the rules in practice; otherwise, their discordance leads to confusion and the possibility of arbitrary state conduct. Fourth, the rule of law places a responsibility on political actors to make the law understandable to those subject to it, and to make known the laws that apply to the governors. The law should be clear so that those subject to it may comply with it.
Dicey and Hayek’s own renderings of the rule of law also suggest that stealth amendment is problematic. Dicey writes that “the rule of law is contrasted with every system of government based on the exercise by persons in authority of wide, arbitrary, or discretionary powers of constraint.” Hayek echoes the same theme in defining the rule of law:
Stripped of all technicalities, [the rule of law] means that government in all its actions is bound by rules fixed and announced beforehand—rules which make it possible to foresee with fair certainty how the authority will use its coercive powers in given circumstances and to plan one’s individual affairs on the basis of this knowledge.
The rule of law requires official conduct to conform to standards established in advance of the actions taken, and it authorizes citizens both to scrutinize that official conduct and to pass judgment upon it.
In contrast to the rule of law’s expectation of limited discretion, the power of stealth amendment expands the discretionary authority of political actors. The exercise of discretion is, of course, not problematic on its own for the rule of law in constitutional democracies. Indeed, discretion is a necessary feature of liberal democratic governance, particularly in light of the rise of the administrative state. But the difference here is that stealth amendment combines the exercise of discretion with informality and irregularity, and together they undermine the values of transparency, accountability, and predictability—three fundamental values that we associate with the rule of law, and which double as operating principles for good government.
In his analysis of the democratic foundations for the rule of law, Joseph Raz highlighted these three values. The rule of law, he wrote, requires transparency: the law should be “open and adequately publicized” because people “must be able to find out what it is.” Laws and their meaning must therefore be clear: “[a]n ambiguous, vague, obscure, or imprecise law is likely to mislead or confuse at least some of those who desire to be guided by it.” The rule of law also requires predictability: laws “should be relatively stable,” and “should not be changed too often.” If the law changes often, Raz cautioned, “people will find it difficult to find out what the law is at any given moment and will be constantly in fear that the law has been changed since they last learnt what it was.” More broadly, the rule of law demands predictability because “people need to know the law not only for short-term decisions … but also for long-term planning,” and they need to be confident in “their knowledge of the content of the law.” The rule of law also demands accountability for political actors and the decisions they make in their official capacity. Law should be general, open, and stable, and should in turn be subject to laws that are themselves general, open, and stable. The value of accountability in the rule of law therefore derives from the expectation that lawmaking itself must “be guided by open, stable, clear, and general rules.” This in turn puts political actors on notice that citizens must be given the capacity to monitor their conduct.
At bottom, then, the rule of law holds that “political power may not be exercised except according to procedures and constraints prescribed by laws which are publicly known.” It “requires all persons, including governmental officials, to obey the laws and be held accountable if they do not,” and insists that “the laws can be changed only through constitutional procedures and may not be nullified or overridden by individual fiat.” In this way, the rule of law binds political actors to clearly disseminated principles and procedures that are transparently revisable. In contrast, the informality, irregularity, and circumvention of stealth amendment is inconsistent with these rule of law values of transparency, predictability, and accountability.
These are procedural values, but they appeal to us for more than their procedural protections. We value them also for how they shape interactions between public institutions and private individuals, protecting the latter from the former. According to Jeremy Waldron, we “radically sell short the idea of the Rule of Law” where we do not recognize the importance of procedure. Conforming lawmaking to norms of generality, publicity, prospectivity, stability, and clarity helps guard against violations of the substantive values that we associate with liberal democracy, namely dignity and liberty. Political actors should accordingly be sanctioned for violating the rule of law
when the norms that are applied by officials do not correspond to the norms that have been made public to the citizens or when officials act on the basis of their own discretion rather than according to norms laid down in advance.
Here, again, we see the importance to the rule of law of the procedural values of transparency, accountability, and predictability—values reflected in procedures that serve substantive democratic purposes.
2. Formal Amendment and the Rule of Law
Formal amendment procedures serve these three rule of law values. Formal amendment telegraphs when and how constitutional change occurs, and it produces legislatively or popularly agreed upon changes that are ultimately inscribed in the constitutional text for all to read and internalize. Pursuing constitutional change via formal amendment performs an educative function in society, both for political actors and for the people, and the textual memorialization of the change satisfies Fuller’s four criteria noted above for respecting the rule of law: it is a formalized process, it reflects congruence between law and practice, it serves the interest of clarity, and it results in public dissemination. This fourth criterion, publicity, is an important but largely underappreciated function of formal amendment, but in it we find the core of the reason why respecting formal amendment procedures is central to the rule of law: “[t]his textual referent, being available and apparent, enables more people to understand the fact that there has been constitutional change and to take note of it than if the change comes informally.” Formal amendment and its textual entrenchment is structured, visible, and overt, not ad hoc and unseeable, the latter being features of constitutional amendment by stealth.
Formal amendment rules are a corrective vehicle authorizing political actors to alter the constitutional text in a public, knowable, and comprehensible process. Their public procedures invite civic engagement when they are invoked, their textual entrenchment makes them knowable and accessible, and their precise requirements are generally accessible enough so as to allow political actors and the people to understand the standard they set for constitutional changes. Formal amendment rules consequently promote predictability “by assuring that any constitutional modifications are predictable, orderly, strictly regulated, and highly supported.” They help foster stability by making it difficult to alter entrenched commitments, which in turn, can moderate the pace of constitutional change. They also serve transparency, insofar as formal amendment is associated with constancy and clarity.
Constitutional amendment by stealth cannot serve any of these three rule of law values. It fails the tests of predictability, transparency, and accountability, because its procedures are not knowable, at least initially, by anyone other than the political actors who choose to pursue it. This complicates the task of holding political actors accountable. Where the informal procedures of a stealth amendment are unknowable, it becomes unlikely that the political actors pursuing the change can, or eventually will, be held accountable for their non-public decisions before it becomes too late—that is, before the political practice has matured into a convention.
Formal and stealth amendment differ in three other ways. First, formal amendment requires a new constitutional rule to survive the complex but clearly enumerated steps of approval and ratification, and if successful, it culminates in a new rule that has been legitimated by the rigours of entrenched amendment procedures. In contrast, the informal process of constitutional amendment by stealth threatens to devalue the words of the constitutional text. As Brannon Denning cautions, “the reliance on informal change can produce a constitutional culture in which people feel less and less bound by the words of the document which supposedly governs them.” Second, in contrast to formal amendment, which requires the participation of a range of political, popular, and institutional actors, stealth amendment excludes opposing political actors and the people from the otherwise deliberative process of constitutional amendment and thus divests both the process and the product of its democratic legitimacy. This is especially problematic where the stealth amendment targets a fundamental feature of the constitutional regime, as is the case here with respect to the Senate. Finally, stealth amendment denies opposing political actors and the people the right to engage in an open debate about constitutional issues of national importance. This is problematic in a constitutional democracy because the right to engage in the formal process of constitutional amendment is, above all, a right to exercise democracy.
Yet one can resist stealth amendment without rejecting all informal constitutional change. There are advantages to pursuing and authorizing constitutional change outside the strictures of formal amendment rules. For one, the instability and relative impermanence of informal amendment can be recast as a virtue: it authorizes political actors to adapt the constitution to changing times and exigencies without the risk of formal amendment failure. In addition, the unwritten informal amendment process fosters dialogic interactions among political, popular, and institutional actors, and these kinds of interactions are socially constructive. An additional benefit involves constitutional contestation: the difficulty of both identifying and defining the content of informal amendment promotes the continuing contestability of constitutional law. Contestability in this respect is arguably valuable because it has the potential to enhance civic participation in elaborating constitutional meaning, and it might moreover promote judicial minimalism, to the extent this is a desirable judicial posture. Informal amendment can therefore entail important benefits. But constitutional amendment by stealth offers none of them because it is a calculated, non-public circumvention of the rules of democratic constitutionalism.
3. Consultative Elections and the Rule of Law
Here, the Conservative government’s recourse to section 44 to amend senator selection by stealth was driven by the difficulty of formal amendment, and its intent to do informally what is impossible formally. The prime minister’s historic appearance before the Special Senate Committee on Senate Reform—the first time the sitting head of government appeared before a Senate committee—underlined his commitment to Senate reform, but also revealed his strategic calculation to proceed by informal rather than formal amendment. His testimony made clear that his objective was eventually to “have an election process where we can consult the population rather than to appoint senators traditionally.” Such a reversal from tradition should not, however, occur through the normal legislative channels he ultimately chose for initiating Senate reform. It should instead occur only through the formal procedures required by the constitution.
In September 2006, the prime minister appeared before the Senate to discuss his government’s bill on Senate term limits. But he also addressed consultative senatorial elections. The prime minister began his testimony by lamenting the repeated failures of Senate reform in Canadian history. As the newly elected prime minister, he suggested that the same thing would not happen under his leadership: “The government is not looking for another report” but “is seeking action,” he said, insisting that he had made a campaign pledge to reform the Senate and that he had “[come] here today to reiterate personally [his] commitment to reform this institution.”
He then explained that the bill on Senate term limits was part of a larger plan to proceed incrementally to reform the Senate. After term limits, the government would create Senate elections: “[a]s yet another step in fulfilling our commitment to make the Senate more effective and more democratic, the government, hopefully this fall, will introduce a bill in the House to create a process to choose elected senators.” That bill, the prime minister emphasized, would “further demonstrate how seriously the government takes the issue of serious Senate reform.” The prime minister saw an elected Senate as important because it would democratize the body, assuage longstanding misgivings from western Canada, and it would also bring Canada in line with modern constitutional democracies, which “virtually all now elect their legislatures.”
In his exchanges with senators on the committee, the prime minister’s plan became clear. He had chosen to pursue incremental change—Senate term limits first, then consultative elections on their own—because piecemeal change allowed Parliament to act unilaterally “without engaging other levels of government in a complex constitutional discussion or amendment process.” The prime minister acknowledged that he could have attempted these changes all at once by launching a process of “comprehensive reform through, in a sense, mega constitutional negotiations.” But he concluded that “[m]y observations over the last 20 years of federal–provincial politics … are such that I do not see comprehensive Senate reform achievable today, except, perhaps, one kind of comprehensive reform—abolition.” It was obvious from his testimony, however, that he preferred reform over abolition, as he stated that “I will be frank in saying that I tend to think of a future Senate in terms of it being an elected body,” and that “[a]nything short of a democratic electoral process would fall short of what we ultimately need on accountability.” But this change would come about informally because “[t]here is no doubt that to change the process in a formal constitutional sense—to making senators elected—would require provincial consent.” And provincial consent would be unachievable on that issue without triggering wholesale constitutional reform that would be doomed from the start.
The ultimate goal, for the prime minister, was a fully elected Senate. To him, Senate term limits were “an interim step of democratization” that would later lead to consultative elections and rebalancing provincial representation in the Senate, both of which the Conservative government would pursue unilaterally in light of the difficulty of formal amendment. As to the imbalance in provincial representation, the prime minister acknowledged that “[i]n the future, we will have to address this problem but at the same time, the government has to choose a staged approach.” Provincial representation “is perhaps the most difficult issue” in Senate reform “and for this reason, the government did not start with this step. The government started first with the terms and secondly with an election process.” For him, reforming provincial representation and the powers of the Senate “must be addressed through a general amending formula, constitutional amendment,” but changes to Senate term limits and creating senatorial elections could be done by Parliament acting unilaterally.
The prime minister was right as a matter of political expedience—it would be easier and more politically profitable for him to change the Senate legislatively than constitutionally—but he was wrong on the constitution. Nevertheless, his long-term plan to reverse engineer senatorial elections was shrewd but democratically illegitimate. The Conservative government was concerned that proposing Senate reform via formal amendment would have inevitably lead to a large-scale constitutional revision—and this effort would have necessarily failed in light of provincial dissensus not unlike what had felled the Meech and Charlottetown Accords years ago. The Conservative government therefore chose instead to pursue its Senate reforms without defining it as a formal amendment. By creating a framework for consultative senatorial elections to generate candidates for prime ministerial nomination to the Senate—which as I have argued above, would have matured from a voluntary prime ministerial practice into a binding prime ministerial convention—the Conservative government would have bound present and future political actors. Present and future provincial political actors would have been coerced by the democracy principle into adopting and subsequently adhering to the framework for consultative senatorial elections because their constituents would have demanded the continuing right to select their senators. And future prime ministers would themselves have been coerced by the democracy principle and prior political practice to follow the initially voluntary prime ministerial practice of nominating winning candidates to the Senate.
Over time, as the practice matured into a convention, the Senate would have become reconstituted in its composition, and as a result of its new composition, in its democratic legitimacy and legislative function. As to its composition, the Senate would have changed from a wholly appointed body to a de facto elected one. Senators would have, as a technical matter, remained appointed to the Senate by the Governor General on the advice of the prime minister. But, functionally, appointment would have occurred at the provincial level in consultative senatorial elections whose outcome would have effectively determined for the prime minister whom to nominate to the Governor General. An effectively elected Senate would have accrued the democratic legitimacy that it lacks today, a transformation in status that could in turn have emboldened senators to take a more active role in the legislative process. Where the appointed Senate now commonly assents to the House of Commons absent rare exceptions, an elected Senate might no longer treat its assent to the House as the simple formality it currently is.
The evolution of Parliament from two elected chambers, each claiming some measure of democratic legitimacy, would have redesigned Canadian parliamentary government in a way not constitutionally achievable without formal amendment. Those changes would have bred anticipated—as well as unintended—legislative, partisan, geographic, and institutional implications. It is beyond the scope of this article to detail the transformative changes caused by an elected Senate. It is generally understood, though, that “[a]n elected Senate would thus constitute a major change that could only be done by means of a constitutional amendment and a full revision of the operation not only of the Senate, but of the government as a whole.” Whatever the consequences of the major changes occasioned by an elected Senate, the key point is that the constitutional transformation would have arisen informally as a result of the Conservative government’s ordinary legislation to create a framework for consultative elections.
Constitutional changes of this magnitude should occur through the formal, public, contestable, and democracy-enhancing and -preserving channels of constitutional amendment. As the Court recognized in the Senate Reference, amending senator selection must be done through the multilateral amendment procedures entrenched in section 38, not the unilateral federal amendment power in section 44. The reason why, however, which the Court did not develop, is that changes to the basic framework of government—like the introduction of consultative senatorial elections—must be legitimated by successfully navigating the intricate procedures of formal amendment, designed to express the informed aggregated choices of political actors in their capacity as responsible representatives of the electorate. Yet, had they been successful, the informal and irregular procedures deployed by the Conservative government would have obscured the reality and extent of its intended constitutional changes. They were insufficiently predictable, they lacked transparency, and they inhibited public accountability.
Entrenching major constitutional reforms by stealth comes at a cost. There are costs even where the attempt to amend the constitution by stealth ultimately fails, as here with the repudiated proposal for consultative senatorial elections. For example, there are political costs to the moral standing of the governing party, and to its trustworthiness as perceived by citizens and opposing political actors when its non-constitutional tactics of stealth amendment are brought to light. But those political costs are not the focus of this section. I am instead interested here in how stealth amendments to national institutions affect parliamentary, provincial, and popular actors in discharging their obligations in the process of constitutional change. Stealth amendment denies them their democratic right to participate in constitutional change, and thereby degrades what is designed to be a public, deliberative, representative, and collaborative process into a closed, arbitrary, unrepresentative, and deeply problematic one. In this section, I explore the costs and consequences of stealth amendment, and evaluate why it is problematic for the rule of law in a constitutional democracy. First, however, I begin by distinguishing stealth amendment from other forms of conventional constitutional change.
1. Stealth Amendment and Ordinary Conventional Change
Scholars of comparative public law might draw similarities between the phenomenon of constitutional amendment by stealth in Canada, and the expansion of presidential war powers in the United States. Indeed, they might argue that the now-common presidential practice of committing the armed forces into conflict abroad without a congressional declaration of war has affected not only what Stephen Griffin has referred to as “an amendment-level change to the constitutional order outside the [formal] amendment process” but more specifically an informal amendment by stealth. This is not an implausible comparison. After all, presidents have routinely deployed troops into combat without seeking congressional approval, in apparent violation of the United States Constitution’s textual command that only “Congress shall have power [t]o … declare War.” And as the practice has persisted, it has created precedents upon which succeeding presidents have relied to legitimate the choice to deploy the armed forces without formal congressional approval. This has informally entrenched the presidential prerogative to circumvent the declaration of war clause, and all of this occurred by stealth. This, at least, would be the nature of the parallel drawn by scholars of comparative public law.
Yet there is something distinctive about constitutional amendment by stealth that makes it an inappropriate category of informal constitutional change into which to classify the informal amendment of presidential war powers. At the origins of stealth amendment are self-consciously undertaken actions to exercise official authority in a manner that will make it politically unpalatable for successors to refuse to conform their conduct to that action and therefore to alter constitutional practice without a new textual writing. The objective of stealth amendment is to impose an unwritten non-legal political obligation on future political actors to follow a certain course of action that they may not necessarily have chosen for themselves absent the constraint forced upon them by a previous political actor who, by deliberate conduct, has narrowed the range of choices successors have in discharging the duties of their office. It is important to stress that the obligation is not a legal one, inasmuch as it is nowhere entrenched in a constitutional or legislative text, nor has it been legitimated by popular or legislative measures. It is a purely political obligation, though its effect approximates a binding legal responsibility.
The modern presidential prerogative to commit the armed forces into combat without a congressional declaration of war may in fact have informally amended the constitution, but it has not occurred stealthily. Presidents have not actively sought to alter constitutional practice, nor can we trace the origin of this presidential prerogative to a self-consciously undertaken decision to act in a manner that would create a binding expectation that successor presidents would have no politically acceptable alternative but to follow the precedent set by the original actor.
Rather, it seems that the presidential prerogative to deploy troops without a congressional declaration of war was born of congressional self-interest in disclaiming responsibility for presidential commitments abroad and to instead push any potential blame to the president. Congress’s refusal to exercise its constitutional power to insist on a declaration of war has left the president with almost unencumbered authority to wage war. And presidents, for their part, “have been willing to accept responsibility for wars … only after Congress thrust it upon them because its members decided that blame avoidance was the winning political strategy.” Today, questions remain about whether congressional declarations of war are a necessary condition of the use of military force abroad. But the president nonetheless continues to engage the armed forces in foreign conflicts without congressional declarations, as he has done since the Second World War—the last time Congress formally declared war. The president does so, however, not because of a political or legal obligation foisted upon him by a predecessor. This informal change in constitutional practice cannot therefore be described as having occurred by stealth.
We may also contrast the creation of the two-term convention on presidential tenure prior to the entrenchment of the Twenty-Second Amendment with the creation of a constitutional amendment by stealth. It is a subtle distinction, but the difference between the introduction of consultative senatorial elections and the emergence of the two-term limit accentuates the essential feature of stealth amendment: intent. Whereas both convention and stealth amendment arise out of political practice over time as a function of continuity, convenience, or even perceived though not deliberate compulsion, we can discern at the point of origin of a stealth amendment the deliberate intent of political actors to create an unwritten norm that will coerce their successors into conforming their conduct to the practice. Such a coercive intent is not present in the normal course of conduct that ultimately becomes informally entrenched as a convention. But it is this intent that defines the core of what I identify as a stealth amendment.
The Twenty-Second Amendment imposes a two-term limit on the president. By its terms,
[n]o person shall be elected to the office of the President more than twice.
The period between its conception and entrenchment was rather short at four years: proposed in 1947 and ratified in 1951, the impetus behind the amendment was the unprecedented fourth term to which President Franklin Delano Roosevelt had been elected in 1944. Although there is truth to the claim that Roosevelt’s Republican adversaries intended the amendment as “a belated slap at him,” and as a deeply partisan expression of dissent to his enacted legislative programs, the larger purpose of the Twenty-Second Amendment was to thwart the rise of an imperial presidency that would accumulate power at the expense of the other branches.
But well before the Twenty-Second Amendment formally bound presidents to serve only two terms, presidents believed themselves bound by that rule, and third parties likewise understood presidents to be bound by it. Politicians and commentators saw the two-term limit as “normatively obligatory, central to the maintenance of the U.S. constitutional project,” and therefore much “[m]ore than just an observed historical pattern.” No court would have enforced the rule, but as Dicey wrote, it possessed “in practice nearly the force of law.” The reason why recalls our analysis of the creation of a constitutional convention. There had emerged over time a convention pursuant to which presidents would not seek a third term. We can trace the origin of this convention to the first president of the United States, George Washington, who chose not to run for reelection after serving two full terms. In declining to seek a third term in 1796, Washington began what was an initially self-policed practice of voluntary presidential resignation that later became accepted as a tradition of informal presidential term limits, as president after president followed the Washington precedent. Scholars have recognized that the Washington precedent was the basis for creating a convention limiting presidents to two consecutive elected terms in office.
The key point for our comparative study of stealth amendment is how the Washington precedent matured into a convention. Washington did not refuse to run for a third consecutive term in order to model the behaviour he intended his successors to follow. He declined to run, although he likely would have won, because he wished to retire to private life, not out of duty, nor a sense of constitutional propriety, nor an intent to coerce future presidents into respecting his two-term tradition. As Bruce Peabody explains, Washington was not “the willful founder of a custom of presidential term limits,” a claim that scholars have commonly made. Indeed, political scientist Paul Davis observes that “there is ample evidence that he never expected or desired his refusal to become a precedent for later presidents.” This did not prevent political actors from pointing to the Washington precedent as a model of selfless leadership to which others should aspire. As each of Washington’s two-term successors from Thomas Jefferson through Andrew Jackson chose one after another not to seek a third consecutive elected term despite there being no textual rule standing in the way, the two-term limit grew by the late nineteenth century into “an unwritten constitutional norm” such that public resistance greeted any president who publicly considered departing from it.
The two-term convention did not arise by stealth. To draw an analogy, “[j]ust as legal precedents acquire their meanings in subsequent decisions,” so too the Washington precedent grew into a convention “primarily in the hands of his successors.” The two-term convention arose as conventions ordinarily do: on the strength of the sustained repetition of the accepted practices of political actors. Over time, succeeding presidents imputed to Washington’s precedent a principled basis that Washington had not himself intended, namely of the importance in rotating the presidency in the service of democracy and of guarding against the concentration of power in a single office. Term limits, they argued and indeed believed, would frustrate the “potential for tyranny,” and would better ensure the “health and vitality” of the president. The democracy principle was therefore only retroactively applied to justify the two-term convention.
Here, there is both contrast and continuity with the stealth amendment unsuccessfully pursued by the Conservative government. As a matter of contrast, the proposed framework of consultative senatorial elections was designed specifically with the intent to compel future prime ministers to follow the model set by the incumbent prime minister. The continued repetition of the practice of senatorial nomination would have become a convention, as future prime ministers followed the precedent intentionally set by their predecessor. The continuity between the stealth amendment and the Washingtonian two-term convention is centred on democratic principle: the same reason that explains why prime ministers would be powerless to depart from the convention of nominating winning consultative senatorial election candidates also explains why Washington’s successors could not themselves depart from his two-term precedent. Just as Washington’s precedent was supported by the principle of democracy—on the theory that term limits prevent the concentration of power—the convention on prime ministerial senate nominations would likewise have grown to be supported by democratic principle; here, the democratic interest of citizens in voting for their senators. The difference, however, is that the Washington precedent was only later imbued with democratic virtue. In contrast, the prime ministerial practice of nominating winning election candidates would have been designed from the beginning to become a convention on the strength of its unimpeachable democratic veneer.
2. Constitutional Integrity and Democratic Legitimacy
But suppose Washington had intended his refusal to run for a third term to be a model for successors, and assume that he had invoked the democracy principle to justify his self-imposed term limit. Further, suppose that Washington’s successors had followed his precedent such that, over time, the two-term limit had matured into a convention, just as Washington had strategically planned. Even this counterfactual wrinkle of Washington’s intent would have been insufficient to classify this new convention as a stealth amendment. This highlights another important criterion for constitutional amendment by stealth: circumvention of the constitutional text.
A stealth amendment occurs where political actors calculate that it is too difficult to formally amend the constitution. Political actors consequently turn their attention to intentionally creating a constitutional convention through repeated political practice. The result is an informal constitutional amendment unlike others: ordinarily, an informal amendment arises with the affirmative approval or acquiescence of political actors engaged in dialogic interactions, and this informal amendment can therefore claim some democratic legitimacy from its origin in political agreement. Although a stealth amendment is a similarly unwritten though binding amendment, it is achieved through an irregular process of compulsion designed to obscure its intended effect until it is too late to deny the democratic legitimacy retrospectively assigned to it.
What made the Conservative government’s proposed framework of consultative senatorial elections irregular is its careful design to circumvent the Constitution of Canada’s formal amendment rules for changing senator selection. As I have explained at greater length elsewhere, the constitution requires political actors to successfully navigate the onerous multilateral amendment procedure in section 38 in order to change the method of selecting senators. This textual command is reinforced by the history and architecture of formal amendment in Canada as well as the spirit of Canadian federalism. Yet the Conservative government sought to make this historic change to Canadian federalism with its narrow and relatively easy (given its parliamentary majority) unilateral formal amendment power in section 44. Cloaking the significance of the change under the cover of a simple statute was intended to convey the impression that the intended change was not as significant as it really was.
The Supreme Court ultimately stopped the Conservative government from amending the constitution by stealth. But had the Court approved the Conservative government’s use of section 44, it would have become a political impossibility, though not a legal one, to remove from citizens the power to select their own senators by the time successor political actors had recognized that they had become compelled to follow the political practice of nominating the winning senatorial election candidates. This practice would have matured into a constitutional convention. Two things would have made this stealth amendment even more troubling: first, the convention would have arisen out of a deliberate effort to circumvent the formal rules of constitutional amendment that prescribe how to create an elected Senate; and second, the convention would have grown to possess democratic legitimacy without having at its point of origin conformed to our expectations for democratic government under the rule of law.
In a constitutional democracy governed by a written constitution with rules for formally amending the document, political actors should abide by the textual rules for constitutional change where the change they seek to effect is governed by a clear rule. Circumventing the constitutional text, as the Conservative government tried to do in its Senate reforms, degrades the constitution and undermines the rule of law. It degrades the constitution by signaling to political actors and the public that the constitutional text does not in fact bind in all cases, and that its authority is contingent on the political preferences of the governing party. It moreover undermines the rule of law for the reasons elaborated above: it fails the tests of predictability, transparency, and accountability.
The act of amending the constitution should reflect the considered judgment of the political community and the popular legitimacy that only deliberative procedures can confer. In the classic Lockean tradition of representative government, a constitutional amendment expresses the consent of the governed, and the granting of its consent legitimizes the amendment its representatives have effected in its name. In this respect, a constitutional amendment is an event of “high moment” in the life of a constitutional democracy insofar as it commonly requires an extraordinary legislative measure, popular agreement, or both. It is “a fundamental act of popular sovereignty,” recourse to which is a reminder to political actors that constitutional legitimacy derives from the direct or mediated consent of public institutions and citizens acting in concert to give meaning to the constitution.
Where one governmental institution—here, the governing majority in the House of Commons—arrogates to itself and denies others the power of constitutional amendment (a collateral consequence of stealth amendment), there is a cost to the Constitution of Canada. Circumventing the textual strictures of the constitution’s formal amendment rules in order to do informally what it commands must be done formally diminishes the integrity of the constitution as it becomes perceived as an ineffective constraint on political actors. There is an equally troubling cost to democracy in Canada where political actors engage in stealth amendment: it divests the practice of amendment of its public, deliberative, opinion-aggregating, democracy-enhancing, and democracy-in-action properties. Stealth amendment therefore has no claim to the democratic legitimacy ordinarily associated with a constitutional amendment, insofar as the only legitimacy a stealth amendment might enjoy is assigned retrospectively as a function of its substantive content alone, rather than both its procedural and substantive merit.
3. Intergenerational Precommitment
What validates a formal amendment is not its content alone, but also the process by which it comes into existence. Where differently constituted majorities overcome the formal barriers to lawful constitutional change, the change itself is validated by two forms of legitimacy, both of which are lacking in stealth amendment. First, the change is validated by the sociological legitimacy of the relevant publics accepting it, either affirmatively or by acquiescence, as justified and deserving of support. Second, the change is validated by the legal legitimacy of satisfying the entrenched standard to create new commitments. Meeting that standard is important to keep fidelity with the binding commitments made by the authoring generation.
Stealth amendment lacks the sociological and legal legitimacy that formal amendments possess by virtue of their successful satisfaction of special legislative or popular thresholds. This void calls into question whether a stealth amendment can properly do what a constitutional amendment is supposed to, which is to bind future generations. A strong reason to accept new political commitments created by a constitutional amendment is that the amendment likely required some measure of supermajority agreement expressed at one or more points in time. There is even greater reason to accept as valid new political commitments where the amendment procedures used to formalize the change are designed to reflect the considered judgments of all parts of the constitutional community. But in the absence of sociological and legal legitimacy, the reasons that commonly justify binding future generations to a constitutional change become less relevant. This intergenerational dimension of constitutional change remains understudied, but it offers a further avenue for understanding the costs and consequences of stealth amendment.
Here, the binding quality of the prime ministerial convention on senatorial nomination would have taken root fraudulently, without having earned the sociological and legal legitimacy that we commonly associate with a constitutional change of that significance. Neither political actors nor the relevant publics acting through their representatives would have been afforded their constitutionally entrenched right to openly express or withhold their freely given consent to this substantial modification to the composition and function of the Parliament of Canada. It would therefore lack sociological legitimacy, just as it would lack legal legitimacy in light of the governing party’s circumvention of the textually prescribed rules for making this change to Canada’s basic constitutional structure. This would in turn deny the convention the democratic authority it needs in order to legitimately bind future generations. Paradoxically, however, the convention would possess independent moral legitimacy anchored in the democracy principle. Notwithstanding the democratically deficient manner in which the convention would have arisen, senatorial elections would have democratic merit and would be worth supporting. This is why stealth amendment must be discouraged: it tempts us to forgive the means in light of the ends. But in constitutional democracy rooted in the rule of law, the means must always be legitimate.
Informal amendment is common in constitutional democracies, including in Canada, where it occurs by judicial interpretation, statutory law, and also by political practice. Stealth amendment is a species of informal amendment but it differs from these conventional forms of informal amendment on one important point: the intent to coerce successors into compliance. Stealth amendment is a deliberate response to the extraordinary difficulty of formally amending the constitution. Recognizing that formal amendment is improbable, political actors circumvent the public, transparent, predictable, and constitutionally required rules for formal amendment and instead proceed through informal and irregular procedures to introduce a material change to the Constitution of Canada. Political actors self-consciously establish a new democratic political practice whose repetition is intended to compel their successors into compliance with that practice.
Notwithstanding whether the new political practice may improve or diminish democratic outcomes, the new practice is not born out of democratic procedures. It is instead an effort to evade the formal, legitimacy-conferring, and democracy-promoting procedures of constitutional amendment that are designed to express the informed aggregated choices of political actors. Over time, this repeated political practice matures into a constitutional convention that becomes informally entrenched in the constitution, yet without the democratic legitimacy we commonly associate with constitutional amendments. This stealth amendment takes root even as political actors convey the impression that no such change is actually occurring.
Informal amendment serves important democratic interests but stealth amendment fails to serve any of them. Constitutional amendment by stealth is an informal yet irregular process of constitutional change that excludes opposing political actors and the people from what is intended to be a deliberative exercise in self-definition. It divests both the process and its eventual product of democratic legitimacy, it denies political actors and the people their fundamental right to democracy, and it moreover degrades what should be a public and collaborative procedure into a closed and coercive one. All of this threatens to devalue the constitutional text and to undermine the rule of law. It is difficult to find any redeeming constitutional virtue in the politically expedient tactic of constitutional amendment by stealth.
The Conservative government tried unsuccessfully to amend senator selection by stealth. In the Senate Reference, the Supreme Court advised the Conservative government that its proposal for introducing consultative senatorial elections would not satisfy the standard the Constitution of Canada sets for constitutional change. Yet in denying the Conservative government its stealth amendment to the Senate, the Court missed an opportunity to bring to light the larger and more fundamental constitutional infirmities with the Conservative government’s Senate reform ambitions. The problem was not that the Conservative government had proposed to introduce consultative senatorial elections. Modernizing the Senate into an elected body is a good idea to consider. Rather, the problem is how the Conservative government had proposed to introduce consultative senatorial elections—by stealth, in violation of the democratic values of transparency, accountability, and predictability in the rule of law.
In this article, I have suggested why we should resist constitutional amendment by stealth. I have explained and illustrated the phenomenon of constitutional amendment by stealth, I have theorized how it emerges in a constitutional democracy, I have posited its interrelation to constitutional rigidity and political impasse, and I have identified its distinguishing features in comparative and theoretical perspectives.
We have not seen the last effort to amend the constitution by stealth. In light of the memorable failures of large-scale constitutional reform in Canada, political actors are unlikely to embark on similarly grand efforts for wholesale constitutional revision. Yet the same problems that led to those failed efforts persist today, unsolved and perhaps as vexing as ever. Incumbents might in the future seek to make incremental progress on Canada’s constitutional challenges by using the constitution’s public, deliberative, and contestable procedures to engage all political actors and the public in some respect. Or they could once again pursue these changes by stealth. Now that we are equipped with a standard and vocabulary to understand constitutional amendment by stealth, we can better monitor political actors who may be tempted to change the Constitution of Canada outside of the legitimate mechanisms of informal amendment and the formal procedures prescribed by the constitutional text.
 1987 Constitutional Accord (Ottawa: Library of Parliament, 3 June 1987) [Meech Lake Accord].
 Consensus Report on the Constitution (Ottawa: Library of Parliament, 28 August 1992) [Charlottetown Accord]. See Ronald L Watts, “Canada: Three Decades of Periodic Federal Crises” (1996) 17:4 Intl Political Science Rev 353 at 355–63.
 Reference Re Secession of Quebec,  2 SCR 217, 161 DLR (4th) 385 [Secession Reference].
 See Peter Oliver, “Canada, Quebec, and Constitutional Amendment” (1999) 49:4 UTLJ 519 at 519.
 Reference Re Senate Reform, 2014 SCC 32, 2014 1 SCR 704 [Senate Reference].
 Constitution Act, 1982, being Schedule B to the Canada Act 1982 (UK), 1982, c 11, Part V [Constitution Act, 1982].
 See Senate Reference, supra note 5 at para 2.
 See ibid at para 5.
 For a description of the various formal amendment procedures and their associated thresholds, see Part I.A, below.
 Richard Albert, “Constructive Unamendability in Canada and the United States” (2014) 67 SCLR (2d) 181 at 218 [Albert, “Constructive Unamendability”].
 Ibid at 213–14.
 Senate Reference, supra note 5 at para 4.
 The Supreme Court has historically answered reference questions expansively, not narrowly, which militates against the view that the Court should have taken such a modest approach in the Senate Reference. See generally James L Huffman & MardiLyn Saathoff, “Advisory Opinions and Canadian Constitutional Development: The Supreme Court’s Reference Jurisdiction” (1990) 74:6 Minn L Rev 1251 (tracing the history of the Supreme Court’s reference jurisdiction and case law).
 CES Franks, “The Senate and its Reform” (1987) 12:3 Queen’s LJ 454 at 454.
 See Robert A Mackay, The Unreformed Senate of Canada (London: Oxford University Press, 1926) at 1.
 See Serge Joyal, “Reflections on the Path to Senate Reform” (1999) 22:3 Can Parliamentary Rev 2 at 2.
 See Michael Crommelin, “Senate Reform: Is the Game Worth the Candle?” (1989) 23:2 UBC L Rev 197 at 197.
 See e.g. George Ross, The Senate of Canada: Its Constitution, Powers and Duties Historically Considered (Toronto: Copp, Clark Company, 1914) at 91–108; Henry S Albinski, “The Canadian Senate: Politics and the Constitution” (1963) 57:2 American Political Science Rev 378 at 378; E Russell Hopkins, “What’s Right about the Senate” (1962) 8:3 McGill LJ 167 at 167–68.
 Paul C Weiler, “Confederation Discontents and Constitutional Reform: The Case of the Second Chamber” (1979) 29 UTLJ 253 at 262.
 See e.g. J Patrick Boyer, Our Scandalous Senate (Toronto: Dundurn, 2014) at 213; Donald J Savoie, “Fix, Don’t Axe, the Senate” (2013) 34:7 Policy Options 7 at 7–8, online: <www.policyoptions.irpp.org/issues/the-age-of-man/savoie/>; Brad Wall, “Time to Consider Abolition of the Senate” (2013) 36:4 Can Parliamentary Rev 6 at 7.
 For recent work on the history of the Senate reform movement, see Bob Plamondon, Full Circle: Death and Resurrection in Canadian Conservative Politics (Toronto: Key Porter Books, 2006) at 51–145; Blair Armitage, A Means to Many Ends: Why Iterative Reform of the Senate is So Difficult (MA Thesis, Carleton University, 2014) at 96–111 [unpublished].
 As a formal matter, senators are appointed by the Governor General on the advice of the prime minister. See Constitution Act, 1867 (UK), 30 & 31 Vict, c 3, s 24 [Constitution Act, 1867].
 See e.g. Aaron Wherry, “Why Wouldn’t the Liberals and NDP Make a Deal to Replace Stephen Harper?”, Maclean’s (19 March 2015), online: <www.macleans.ca/politics/why-wouldnt-the-liberals-and-ndp-cooperate-to-replace-stephen-harper/>; Pierre Saint-Arnaud, “NDP Leader Tom Mulcair Remains Open to Liberal Alliance to Knock Out Tories”,
Toronto Star (18 March 2015), online: <www.thestar.com/news/canada/2015/03/18/ndp-leader-tom-mulcair-remains-open-to-liberal-alliance-to-knock-out-tories.html>.
 As I will suggest in this article, process and purpose must each meet the test of legitimacy in order for a constitutional change to satisfy the requirements of democratic constitutionalism. Although this stealth senatorial amendment is not as deeply problematic as the formally constitutional changes we see occurring at the constitutional and sub-constitutional levels around the world, this stealth amendment does reflect the same kind of general problem that risks undermining constitutional democracy. For two excellent papers on the use of democratic procedures to achieve non-democratic ends, see David Landau, “Abusive Constitutionalism” (2013) 47:1 UC Davis L Rev 189; Ozan O Varol, “Stealth Authoritarianism” (2015) 100:4 Iowa L Rev 1673.
 See John W Burgess, Political Science and Comparative Constitutional Law, vol 1: Sovereignty and Liberty (Boston: Ginn & Company, 1893) at 137.
 The functional uses of formal amendment rules in sham constitutional regimes are much less clear. For a study of the forms and uses of constitutionalism in authoritarian constitutions, see David S Law & Mila Versteeg, “Constitutional Variation among Strains of Authoritarianism” in Tom Ginsburg & Alberto Simpser, eds, Constitutions in Authoritarian Regimes (New York: Cambridge University Press, 2014) 165.
 See Rosalind Dixon & Richard Holden, “Constitutional Amendment Rules: The Denominator Problem” in Tom Ginsburg, ed, Comparative Constitutional Design (New York: Cambridge University Press, 2012) 195.
 See Brannon P Denning & John R Vile, “The Relevance of Constitutional Amendments: A Response to David Strauss” (2002) 77:1 Tul L Rev 247 at 275.
 See András Sajó, Limiting Government: An Introduction to Constitutionalism (Budapest: Central European University Press, 1999) at 39–40.
 See Raymond Ku, “Consensus of the Governed: The Legitimacy of Constitutional Change” (1995) 64:2 Fordham L Rev 535 at 571.
 See Walter Dellinger, “The Legitimacy of Constitutional Change: Rethinking the Amendment Process” (1983) 97:2 Harv L Rev 386 at 431.
 See Jon Elster, Ulysses Unbound: Studies in Rationality, Precommitment, and Constraints (Cambridge: Cambridge University Press, 2000) at 101–104.
 See Rosalind Dixon, “Constitutional Amendment Rules: A Comparative Perspective” in Tom Ginsburg & Rosalind Dixon, eds, Comparative Constitutional Law (Cheltenham, UK: Edward Elgar, 2011) 96 at 97.
 See Richard Albert, “The Expressive Function of Constitutional Amendment Rules” (2013) 59:2 McGill LJ 225 at 236.
 See Richard Albert, “The Structure of Constitutional Amendment Rules” (2014) 49:4 Wake Forest L Rev 913 at 944–46 [Albert, “Structure”].
 See Richard Albert, “Amending Constitutional Amendment Rules” 13 Intl J Constitutional L [forthcoming in 2015].
 See Richard Albert, “How Unwritten Constitutional Norms Change Written Constitutions” 39 Dublin University LJ [forthcoming in 2015].
 See Richard Albert, “Constitutional Amendment by Constitutional Desuetude” (2014) 62:3 Am J Comp L 641 at 656–69.
 See Richard Albert, “The Theory and Doctrine of Unconstitutional Constitutional Amendment in Canada” [unpublished].
 See Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries (New Haven: Yale University Press, 1999) at 220.
 Elsewhere, I explain in detail why formal amendment in Canada is harder than the constitutional text suggests. See Richard Albert, “The Difficulty of Constitutional Amendment in Canada” 53 Alta L Rev [forthcoming in 2015] [Albert, “Difficulty of Constitutional Amendment”].
 See Allan C Hutchinson, “Constitutional Change and Constitutional Amendment: A Canadian Conundrum” in Xenophon Contiades, ed, Engineering Constitutional Change: A Comparative Perspective on Europe, Canada and the USA (Abingdon, UK: Routledge, 2013) 51 at 57–70.
 See Tom Ginsburg & James Melton, “Cultures of Amendment and the Challenges of Measuring Amendment Difficulty” 13 Intl J Constitutional L [forthcoming in 2015].
 See Albert, “Structure”, supra note 36 at 918–28.
 See Astrid Lorenz, “How to Measure Constitutional Rigidity: Four Concepts and Two Alternatives” (2005) 17:3 J Theoretical Politics 339 at 359.
 See Edward Schneier, Crafting Constitutional Democracies: The Politics of Institutional Design (Oxford: Rowman & Littlefield, 2006) at 224.
 See Donald S Lutz, Principles of Constitutional Design (New York: Cambridge University Press, 2006) at 179, n 16.
 Walter Dellinger, “The Amending Process in Canada and the United States: A Comparative Perspective” (1983) 45:4 Law & Contemp Probs 283 at 297.
 See Zachary Elkins, Tom Ginsburg & James Melton, The Endurance of National Constitutions (New York: Cambridge University Press, 2009) at 101.
 Part I of the Constitution Act, 1982, being Schedule B to the Canada Act 1982 (UK), 1982, c 11 [Charter]. See also Edward McWhinney, “The Constitutional Patriation Project, 1980–82” (1984) 32:2 Am J Comp L 241 at 252.
 Michael B Stein, “Canadian Constitutional Reform, 1927–1982: A Comparative Case Analysis Over Time” (1984) 14 Publius 121 at 139.
 “Constitutional Paralysis: Why Canadian Constitutional Initiatives are Doomed to Fail” (1994) 27:4 Can J Political Science 747 at 748.
 See ibid.
 See ibid.
 See ibid. For example, the Charlottetown Accord satisfied Quebec’s demand for a recognition of its distinctiveness but this undermined the western provinces’ demand for provincial equality—a demand that the Charlottetown Accord entrenched by giving all provinces a veto over constitutional amendments and also by making side payments to western Canada as additional compensation (see Lusztig, supra note 53 at 761).
 See Christopher P Manfredi & Michael Lusztig, “Why Do Formal Amendments Fail? An Institutional Design Analysis” (1998) 50 World Politics 377 at 380. See also David R Cameron & Jacqueline D Krikorian, “Recognizing Quebec in the Constitution of Canada: Using the Bilateral Constitutional Amendment Process” (2008) 58 UTLJ 389 at 394.
 See Manfredi & Lusztig, supra note 57 at 399.
 Note that Parliament also possesses a narrow power of amendment under section 101 of the Constitution Act, 1867.
 See Senate Reference, supra note 5 at para 5.
 Constitution Act, 1982, supra note 6, s 45.
 Ibid, s 44.
 See ibid.
 Ibid, s 43. Although the use of section 43 in connection with funding to public religious schools is not specified in the constitutional text, it has been upheld by courts in both Newfoundland and Quebec. See Hogan v Newfoundland (AG), 2000 NFCA 12, 183 DLR (4th) 225, leave to appeal to SCC refused, 27865 (9 November 2000); Potter v Quebec (AG),  RJQ 2823, 109 ACWS (3d) 716 (CA), leave to appeal to SCC refused, 29016 (31 October 2002).
 See Constitution Act, 1982, supra note 6, s 43.
 See ibid, s 38.
 See ibid, s 38(1).
 See ibid.
 See ibid, s 42(1).
 See ibid, s 38.
 See ibid, s 41.
 See ibid.
 See ibid.
 Bettina Petersohn, “Constitutional Reform and Federal Dynamics: Causes and Effects” in Arthur Benz & Jörg Broschek, eds, Federal Dynamics: Continuity, Change, and the Varieties of Federalism (Oxford: Oxford University Press, 2013) 297 at 316.
 See Peter W Hogg, Constitutional Law of Canada, 5th ed (Toronto: Carswell, 2007) (loose-leaf 2014 supplement), 1-7–1-8, n 32.
 See Ian Greene, “Constitutional Amendment in Canada and the United States” in Stephen L Newman, ed, Constitutional Politics in Canada and the United States (Albany: State University of New York Press, 2004) 249 at 254.
 See Fair Representation Act, SC 2011, c 26.
 Stephen M Griffin, “The Nominee is … Article V” (1995) 12 Const Commentary 171 at 172.
 See Heather K Gerken, “The Hydraulics of Constitutional Reform: A Skeptical Response to Our Undemocratic Constitution” (2007) 55 Drake L Rev 925 at 929.
 See ibid at 927.
 See Richard Albert, “Constitutional Disuse or Desuetude: The Case of Article V” (2014) 94 BUL Rev 1029 at 1060–77.
 See Richard A Primus, “When Should Original Meanings Matter?” (2008) 107:2 Mich L Rev 165 at 210, n 158.
 Supra note 43 at 57.
 Ibid at 56.
 Ibid at 57.
 Ibid at 61.
 Ibid at 62 [internal quotation omitted].
  1 SCR 1075, 70 DLR (4th) 385.
 Hutchinson, supra note 43 at 67.
 Ibid at 61.
 In some instances, judicial interpretation is never final given that the Court’s interpretation may be effectively overturned, at least temporarily but theoretically indefinitely, by recourse to the notwithstanding clause (see Charter, supra note 51, s 33). But the clause has become largely inoperable (see Richard Albert, “Advisory Review: The Reincarnation of the Nothwithstanding Clause” (2008) 45:4 Alta L Rev 1037 at 1052–54).
 See Tsvi Kahana, “Canada” in Dawn Oliver & Carlo Fusaro, eds, How Constitutions Change: A Comparative Study (Oxford: Hart, 2011) 9 at 33.
 Reference Re Supreme Court Act, ss 5 and 6, 2014 SCC 21 at paras 90–105,  1 SCR 433 [Nadon Reference].
 Secession Reference, supra note 3 at paras 88–105.
 Indeed, in an important analysis of the Secession Reference, Donna Greschner asks whether the principles the Court recognized as binding on political actors in connection with secession can “contradict or override the written rules of the constitution?” She answers, correctly in my view, that “the opinion’s message is that principles are more important than rules, notwithstanding the pronouncements about the primacy of the text” (“The Quebec Secession Reference: Goodbye to Part V?” (1998) 10:1 Const Forum Const 19 at 23).
 Whether a formal amendment changes the essential features of the Court will of course depend on how the Court interprets that amendment; specifically, whether it requires conformity with the general default or the unanimity amendment procedure.
 Nadon Reference, supra note 94 at para 94.
 See ibid at paras 90–105.
 An example of a parliamentary law that has informally amended Canada’s formal amendment rules is the Clarity Act. See An Act to give effect to the requirement for clarity as set out in the opinion of the Supreme Court of Canada in the Quebec Secession Reference, SC 2000, c 26 (“Clarity Act”). I discuss the effect of the Clarity Act on Canada’s formal amendment rules in Albert, “Difficulty of Constitutional Amendment”, supra note 42.
 See An Act Respecting Constitutional Amendments, SC 1996, c 1, s 1.
 See Robert A Young, “Jean Chrétien’s Québec Legacy: Coasting Then Stickhandling Hard” (2004) 9:1 Rev Const Stud 31 at 38–39.
 See Andrew Heard & Tim Swartz, “The Regional Veto Formula and Its Effects on Canada’s Constitutional Amendment Process” (1997) 30:2 Can J Political Science 339 at 342–43.
 See David E Smith, “The Canadian Senate: What Is To Be Done?” in Christian Leuprecht & Peter H Russell, eds, Essential Readings in Canadian Constitutional Politics (Toronto: University of Toronto Press, 2011) 43 at 43.
 See CES Franks, “A Continuing Canadian Conundrum: The Role of Parliament in Questions of National Unity and the Processes of Amending the Constitution” in
J Peter Meekison, Hamish Telford & Harvey Lazar, eds, Reconsidering the Institution of Canadian Federalism (Montréal: McGill-Queen’s University Press, 2004) 35 at 44.
 I disagree with this conventional view because we cannot yet be certain whether such a constitutional convention requiring referendal consultation has yet taken root. See Richard Albert, “The Conventions of Constitutional Amendment in Canada: Is There a Convention of Federal Referendal Consultation?” (Paper delivered at the Symposium on Constitution-Making and Constitutional Design, Boston College Law School, 31 October 2015).
 See e.g. Katherine Swinton, “Amending the Canadian Constitution: Lessons from Meech Lake” (1992) 42 UTLJ 139 at 167–68 (1992) (arguing that formal amendment should now incorporate greater citizen participation); R Kent Weaver, “Political Institutions and Conflict Management in Canada” (1995) 538 Annals Am Acad Pol & Soc Sci 54 at 65.
 See Chief Electoral Officer of Canada, The 1992 Federal Referendum: A Challenge Met (Ottawa: Chief Electoral Officer of Canada, 17 January 1994) at 58, online: <www.
elections.ca/content.aspx?section=res&dir=rep/off/1992&document=index&lang=e> (cataloguing provincial, territorial, and national participation rates in the 1992 referendum).
 I discuss in great detail each of these extra-textual sources of formal amendment difficulty in Albert, “Difficulty of Constitutional Amendment”, supra note 42.
 Modern proposals for Senate reform have generally called for a “Triple-E” Senate that is equal, elected, and effective. See Ronald L Watts, “Bicameralism in Federal Parliamentary Systems” in Serge Joyal, ed, Protecting Canadian Democracy: The Senate You Never Knew (Montréal: McGill-Queen’s University Press, 2003) 67 at 96–100; Michael Lusztig, “Federalism and Institutional Design: The Perils and Politics of a Triple-E Senate in Canada” (1995) 25:1 Publius 35 at 36, 39.
 See Our Undemocratic Constitution: Where the Constitution Goes Wrong (And How We The People Can Correct It) (New York: Oxford University Press, 2006) at 29.
 See Reference Re Resolution to Amend the Constitution,  1 SCR 753 at 904–05, (sub nom Reference Re Amendment of the Constitution of Canada (Nos 1, 2 and 3)) 125 DLR (3d) 1.
 See Constitution Act, 1982, supra note 6, Part V.
 For a concise overview of Senate reform proposals, see Canada, Library of Parliament, Reforming the Senate of Canada: Frequently Asked Questions (Background Paper), by Andre Barnes et al, Publication No 2011-83-E (Ottawa: Library of Parliament, 12 September 2011) at 7–12, online: <www.parl.gc.ca/Content/LOP/ResearchPublications/
 Compare E Russell Hopkins, “What’s Right about the Senate” (1962) 8:3 McGill LJ 167 (“[e]ver since Confederation, the Senate has been subject to sporadic attacks—not always or even usually well-informed—on its organization, functions and functioning” at 167) with MacKay, supra note 16 (“[p]robably on no other public question in Canada has there been such unanimity of opinion as on that of the necessity for Senate reform” at 206) and Natalie Stechyson, “More than One-Third of Canadians Say Senate Should Be Abolished: Poll”, National Post (18 February 2013), online: <news.nationalpost.
com/news/canada/canadian-politics/canadian-senate> (reporting that forty-two per cent of Canadians support Senate reform and a total of seventy-eight per cent support either reform or abolition).
 See David E Smith, The Canadian Senate in Bicameral Perspective (Toronto: University of Toronto Press, 2003) at 157–58 (observing that “[t]he great expectations associated with large-scale institutional change are as inappropriate as they are unrealizable”).
 See Bruce M Hicks, “Can a Middle Ground be Found on Senate Numbers?” (2007) 16:1 Const Forum Const 21 at 21.
 The reference questions are archived electronically by the Privy Council Office. See “PC 2013-0070” (1 February 2013), online: Government of Canada <www.pco-bcp.gc.ca>. The three bills concern senator selection, Senate term limits, and consultative Senate elections.
 Bill S-4, An Act to Amend the Constitution Act, 1867 (Senate tenure), 1st Sess, 39th Parl, 2006 (first reading 30 May 2006) [Senate Term Limits bill].
 Ibid, cl 2. Those senators holding a seat prior to the effective date of the bill would retain their seat under the current rules, which authorize a Senator to serve until age 75. See Constitution Act, 1867, supra note 23, s 29.
 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) [Senate Appointment Consultations bill].
 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, cls 4–5 (first reading 21 June 2011) [Senate Reform bill].
 See ibid, cls 2–3.
 Ibid at Schedule 1, Part 1, s 1.
 Ibid at Schedule 1, Part 1, s 2.
 Ibid, cl 3.
 See ibid at Schedule 1, Part 1, s 3.
 See ibid at Schedule 1, Part 1, s 5.
 See ibid at Schedule 1, Part 1, s 7.
 See ibid at Schedule 1, Part 1, ss 8–18.
 See ibid at Schedule 1, Part 1, ss 19–22.
 See ibid at Schedule 1, Parts 2–3.
 See ibid at Schedule 1, Part 1, ss 24–29.
 See ibid, cls 4–5.
 See Constitution Act, 1867, supra note 23, s 29, which confers life tenure, up to age seventy-five, upon senatorial appointees.
 The bill creates rules for administering consultative elections. See Senate Appointment Consultations bill, supra note 121 at Parts 1–2. The bill also sets rules in relation to nominees (see ibid at Part 3), voting and tabulation (see ibid at Parts 4–5), communications and advertising (see ibid at Parts 6–7), as well as finance and enforcement (see ibid at Parts 8–9).
 Ibid at Preamble.
 Ibid (stating that “the power to summon Canadians to the Senate from time to time in the Queen’s name is vested in the Governor General”).
 The difference lies in formal entrenchment: creating direct senatorial elections via section 38 leads to a formal amendment entrenched in the constitutional text. But consultative senatorial elections would not lead to a formal amendment entrenching direct senatorial elections, though it would ultimately create a de facto elected Senate.
 See Senate Appointment Consultations bill, supra note 121 at Part 10.
 Senate Reference, supra note 5 at para 49.
 Ibid at para 70.
 Ibid at para 68.
 See ibid at para 51.
 See supra note 23, ss 24, 32.
 Senate Reference, supra note 5 at para 52.
 Ibid at para 54.
 See ibid at paras 54–61.
 Constitution Act, 1982, supra note 6, s 42(1)(b).
 Senate Reference, supra note 5 at para 64.
 Ibid at para 65.
 See ibid.
 See ibid.
 See ibid at para 51.
 Constitution Act, 1982, supra note 6, s 44.
 Senate Reference, supra note 5 at para 68.
 Ibid at para 69.
 See ibid.
 See ibid.
 See ibid at para 70.
 See ibid at para 62.
 The Canadian Bar Association, National Constitutional and Human Rights Section, “Bill C-20: Senate Appointment Consultations Act” (April 2008) at 5, online: <www.cba.
 Ibid at 8.
 Ibid at 9.
 See “Constitutional Conventions and Senate Reform” (2013) 5 Revue québécoise de droit constitutionnel 107 at 120–21.
 See ibid.
 See “La modification par convention constitutionnelle du mode de désignation des sénateurs canadiens” (2008–09) 39:1–2 RDUS 115 at 141–43.
 See “The Impact of Senate Reform on the Functioning of Committees” (2007) 30:1 Can Parliamentary Rev 16 at 17–18.
 Mark D Walters, “The Constitutional Form and Reform of the Senate: Thoughts on the Constitutionality of Bill C-7” (2013) 7 JPPL 37 at 39 [emphasis in original].
 See ibid at 47.
 Ibid at 48. Here, Walters is drawing from Hart’s distinction between being obliged and having an obligation. Walters appears to have attributed a psychological feeling to both feeling “obliged” to do something and feeling “obligated” to do that thing, although Hart seems to have intended to indicate that being obliged turned on a psychological feeling of compulsion whereas having an obligation was a strictly normative matter. See HLA Hart, The Concept of Law, 2nd ed (Oxford: Oxford University Press, 1994) at 82–84.
 Walters, supra note 173 at 47–48.
 Ibid at 47.
 Ibid [emphasis in original].
 Ibid at 47–48.
 Ibid at 49.
 See ibid.
 See ibid at 50.
 Ibid [emphasis in original].
 See Part III.B, below.
 See Part III.B.3, below.
 Supra note 173 at 49.
 Andrew Heard, Canadian Constitutional Conventions: The Marriage of Law and Politics, 2nd ed (Don Mills, Ont: Oxford University Press, 2013) at 5.
 See ibid at 6–11.
 See The Law and the Constitution, 5th ed (London: University of London Press, 1959) at 136.
 Ibid at 134 [emphasis in original].
 Ibid at 135.
 Ibid at 136.
 Heard, supra note 188 at 7.
 Ibid at 7, 8.
 Ibid at 9.
 Ibid at 11.
 See ibid at 144.
 The Parliament of Canada maintains a list of Senate vacancies by province. As of 11 May 2015, there were six vacancies in Ontario, five in Quebec, three in Manitoba, two each in New Brunswick and Nova Scotia, and one each in British Columbia and Prince Edward Island. See “Standings in the Senate”, online: Parliament of Canada <www.parl.gc.ca/SenatorsBio/standings_senate.aspx?Language=E >.
 See Rosemary Barton, “Stephen Harper under Pressure to Fill Senate Vacancies”, CBC News (8 July 2014), online: <www.cbc.ca/news/politics/stephen-harper-under-pressure-to-fill-senate-vacancies-1.2700273 >.
 See Heard, supra note 188 at 141.
 See Constitution Act, 1867, supra note 23, s 24.
 Hart, supra note 175 at 94.
 Ibid at 102.
 Geoffrey Marshall, Constitutional Conventions: The Rules and Forms of Political Accountability (Oxford: Clarendon Press, 1984) at 217.
 See Jennings, supra note 190 at 136.
 Joseph Jaconelli, “The Nature of Constitutional Convention” (1999) 19:1 LS 24 at 27 [Jaconelli, “Nature”].
 Ibid at 30.
 See Part III.B, below.
 Jaconelli, “Nature”, supra note 209 at 42.
 See Joseph Jaconelli, “Do Constitutional Conventions Bind?” (2005) 64:1 Cambridge LJ 149 at 171–73.
 See Joan Bryden, “Senate Reform or Abolition Unlikely Despite Scandal, Experts Say”, The Globe and Mail (22 December 2013), online: <www.theglobeandmail.com/news/
Chris Hannay, “Where Ottawa and the Provinces Stand on Senate Reform”, The Globe and Mail (13 November 2013), online: <www.theglobeandmail.com/news/politics/where-ottawa-and-the-provinces-stand-on-senate-reform/article15433168/>; Leslie MacKinnon, “Senate Reform Hearings See Provinces Counter Ottawa”, CBC News (12 November 2013), online: <www.cbc.ca/news/politics/senate-reform-hearings-see-provinces-counter-ottawa-1.2423350>.
 See Senate Reference, supra note 5 (Factum of the Attorney General of Canada at paras 130–35).
 See Part III.A.3, above.
 British Columbia v Imperial Tobacco Canada Ltd, 2005 SCC 49 at para 67,  2 SCR 473.
 See ibid at para 58.
 Ibid at para 67.
 Lon L Fuller, The Morality of Law (New Haven: Yale University Press, 1964) at 39.
 Roscoe Pound spoke in this respect of the difference between the law in the books and the law in action (see generally “Law in Books and Law in Action” (1910) 44 Am L Rev 12).
 See Fuller, supra note 220 at 39.
 AV Dicey, Introduction to the Study of the Law of the Constitution, 8th ed (Indianapolis: Liberty Fund, 1982) at 110.
 Friedrich A Hayek, The Road to Serfdom (Chicago: University of Chicago Press, 1944) at 72.
 See Jeremy Waldron, The Rule of Law and the Measure of Property (New York: Cambridge University Press, 2012) at 6.
 In his important critique of Dicey and Hayek, Harry Jones asks, “[i]s discretion such deadly poison to the rule of law that it is better to abandon deeply desired legislative objectives than to run the risk of possible arbitrary use of discretionary power?” He answers no, illustrating why with reference to the challenges of the welfare state. See Harry W Jones, “The Rule of Law and the Welfare State” (1958) 58:2 Colum L Rev 143 at 152.
 See The Authority of Law: Essays on Law and Morality (New York: Oxford University Press, 1979) at 214–19.
 Ibid at 214.
 Ibid at 214–15.
 Ibid at 215, n 5 [emphasis omitted].
 Ibid at 215.
 Fred D Miller, Jr, “The Rule of Law in Ancient Greek Thought” in Mortimer Sellers & Tadeusz Tomaszewski, eds, The Rule of Law in Comparative Perspective (Dordrecht, Netherlands: Springer, 2010) 11 at 11.
 See Jeremy Waldron, “The Rule of Law and the Importance of Procedure” in James E Fleming, ed, Getting to the Rule of Law (New York: New York University Press, 2011) 3 at 4–5.
 Ibid at 6.
 See ibid.
 Ibid at 21. Waldron resists characterizing the rule of law as a wholly “command-and-control aspect of law, or the norm-and-guidance aspect of law,” preferring to stress that the rule of law can and does make room for arguing over the underlying norms that are to bind political actors and citizens (ibid at 22). Waldron therefore urges us to recognize that the rule of law tolerates that “we argue over them adversarially, we use our sense of what is at stake in their application to license a continual process of argument back and forth, and we engage in elaborate interpretive exercises about what it means to apply them faithfully as a system to the cases that come before us” (ibid).
 See Dixon, supra note 34 at 97–98.
 See the text accompanying notes 218–22.
 Denning & Vile, supra note 29 at 279.
 Jane Pek, “Things Better Left Unwritten? Constitutional Text and the Rule of Law” (2008) 83:6 NYUL Rev 1979 at 2009.
 See Bjørn Erik Rasch & Roger D Congleton, “Amendment Procedures and Constitutional Stability” in Roger D Congleton & Birgitta Swedenborg, eds, Democratic Constitutional Design and Public Policy: Analysis and Evidence (Cambridge, Mass: MIT Press, 2006) 319 at 325–26.
 See William E Scheuerman, “Constitutionalism in an Age of Speed” (2002) 19:2 Const Commentary 353 at 378.
 See Brannon P Denning, “Means to Amend: Theories of Constitutional Change” (1997) 65:1 Tenn L Rev 155 at 237.
 See ibid.
 See ibid at 242.
 See Simeon CR McIntosh, Fundamental Rights and Democratic Governance: Essays in Caribbean Jurisprudence (Kingston, Jamaica: Caribbean Law, 2005) at 74.
 See e.g. Xenophon Contiades & Alkmene Fotiadou, “Models of Constitutional Change” in Xenophon Contiades, ed, Engineering Constitutional Change: A Comparative Perspective on Europe, Canada and the USA (Abingdon, UK: Routledge, 2013) 417 at 442.
 See Gerken, supra note 79 at 934.
 See ibid at 937–41.
 See ibid at 941. See also Reva B Siegel, “Constitutional Culture, Social Movement Conflict and Constitutional Change: The Case of the de facto ERA” (2006) 94:5 Cal L Rev 1323 at 1328.
 See Gerken, supra note 79 at 941. For a discussion of judicial minimalism, see generally Cass R Sunstein, One Case at a Time: Judicial Minimalism on the Supreme Court (Cambridge, Mass: Harvard University Press, 1999) (describing and defending the theory of judicial minimalism, which encourages judges to resolve matters on narrow grounds so as to promote democratic deliberation).
 Special Senate Committee on Senate Reform, Proceedings of the Special Senate Committee on Senate Reform (7 September 2006) (Chair: Daniel Hays), online: Parliament of Canada <www.parl.gc.ca/Content/SEN/Committee/391/refo/02ev-e.htm?
Language=EParl> (comments of Rt Hon Stephen Harper, PC, MP, Prime Minister of Canada).
 See ibid.
 See Part III.A, above.
 See “The Supremes Save Canada—and Harper—From a Much Worse Senate”, Editorial, The Globe and Mail (25 April 2014), online: <www.theglobeandmail.com/globe-debate/editorials/the-supremes-save-canada-and-harper-from-a-much-worse-senate/article18231399/>.
 See Jeffrey Simpson, “Why Would We Want an Elected Senate?”, The Globe and Mail (13 February 2013), online: <www.theglobeandmail.com/globe-debate/why-would-we-want-an-elected-senate/article8508794/>.
 See Kathryn Blaze Carlson, “Harper’s Senate Reform Plans May Forever Change How Canada is Governed”, National Post (6 January 2012), online: <news.nationalpost.
 For a discussion of the consequences of an elected Senate, see CES Franks, The Parliament of Canada (Toronto: University of Toronto Press, 1987) at 198–201; Richard Janda, “Canada’s Failed Senate Reform Initiative: On Legitimate and Illegitimate Compromise” (1994) 5 Windsor Rev Legal Soc Issues 1 at 16–39; Peter McCormick, “Could We, Should We, Reform the Senate and the Supreme Court?” (January–February 2000) Policy Options 7 at 7–8, 10; Jack Stilborn, “Forty Years of Not Reforming the Senate: Taking Stock” in Serge Joyal, ed, Protecting Canadian Democracy: The Senate You Never Knew (Montréal: McGill-Queen’s University Press, 2003) 31 at 31–66.
 Gil Rémillard with the collaboration of Andrew Turner, “Senate Reform: Back to Basics” in Serge Joyal, ed, Protecting Canadian Democracy: The Senate You Never Knew (Montréal: McGill-Queen’s University Press, 2003) 105 at 121.
 See Senate Reference, supra note 5 at para 65.
 Stephen M Griffin, Long Wars and the Constitution (Cambridge, Mass: Harvard University Press, 2013) at 30.
 See David Gray Adler, “The Law: George Bush as Commander in Chief: Toward a Nether World of Constitutionalism” (2006) 36:3 Presidential Studies Q 525 at 530.
 US Const art I, § 8.
 See Henry P Monaghan, “Presidential War-Making” (1970) 50 BUL Rev 19 at 31.
 See Daryl J Levinson, “Empire-Building Government in Constitutional Law” (2005) 118:3 Harv L Rev 915 at 955.
 See ibid.
 Ibid at 956.
 See Andrea Shalal, “U.S. Lawmakers Urge Congressional Action to Back Obama’s Syria War”, Reuters (28 September 2014), online: <www.reuters.com/article/2014/09/28/us-syria-crisis-obama-congress-idUSKCN0HN0TO20140928>.
 See Curtis A Bradley & Jack L Goldsmith, “Congressional Authorization and the War on Terrorism” (2005) 118:7 Harv L Rev 2047 at 2060.
 There may be, however, an important parallel in another area of foreign affairs. Powerful nation-states may sometimes undertake a conscious effort to create a new rule of customary international law so as to coerce other nation-states into compliance. See Michael Byers, Custom, Power and the Power of Rules: International Relations and Customary International Law (Cambridge: Cambridge University Press, 1999) at 90–92. This line of inquiry is worth pursuing to elaborate the theory of stealth amendment in international law.
 US Const amend XXII.
 See Richard Albert, “The Evolving Vice Presidency” (2005) 78:4 Temp L Rev 811 at 853.
 Paul H Appleby, “Roosevelt’s Third-Term Decision” (1952) 46:3 American Political Science Rev 754 at 754.
 See Lawrence Schlam, “Legislative Term Limitation Under A ‘Limited’ Popular Initiative Provision?” (1993) 14:1 N Ill UL Rev 1 at 32. See also Paul G Willis & George L Willis, “The Politics of the Twenty-Second Amendment” (1952) 5:3 Western Political Q 469 at 481–82 (noting the “strong political overtones” to the ratification of the amendment).
 See David E Kyvig, “Refining or Resisting Modern Government? The Balanced Budget Amendment to the U.S. Constitution” (1995) 28:2 Akron L Rev 97 at 103.
 Keith E Whittington, “The Status of Unwritten Constitutional Conventions in the United States” (2013) 2013:5 U Ill L Rev 1847 at 1867–68.
 Dicey, supra note 225 at cxliv.
 The convention was more narrow than this: it was that a president would not seek a third or subsequent consecutive elected term unless there were an emergency requiring continuity in the presidency (see Jaconelli, “Nature”, supra note 209 at 33). On this reading, President Franklin Delano Roosevelt did not violate the convention by running for a third and fourth term during the Second World War. Nor did President Theodore Roosevelt violate the convention when he ran for a third term in 1912 because, although he had served almost two full terms, he had been only once elected, having acceded to the presidency in his capacity as vice president as a result of the assassination of William McKinley. For his part, Ulysses S Grant ran unsuccessfully for a third term but it was not consecutive. See Harry A Bailey, Jr, “Presidential Tenure and the Two-Term Tradition” (1972) 2:2 Publius 95 at 98–105.
 See James L Sundquist, Constitutional Reform and Effective Government, revised ed (Washington: Brookings Institution, 1992) at 46.
 See Gideon Maltz, “The Case for Presidential Term Limits” (2007) 18:1 J Democracy 128 at 130–31.
 See e.g. Akhil Reed Amar, America’s Unwritten Constitution: The Precedents and Principles We Live By (New York: Basic Books, 2012) at 462; David E Kyvig, Explicit and Authentic Acts: Amending the U.S. Constitution 1776–1995 (Lawrence, KS: University Press of Kansas, 1996) at 325; Edward S Corwin, The President: Office and Powers 1787–1957: History and Analysis of Practice and Opinion, 4th ed (New York: New York University Press, 1957) at 34.
 See Bruce G Peabody & Scott E Gant, “The Twice and Future President: Constitutional Interstices and the Twenty-Second Amendment” (1999) 83:3 Minn L Rev 565 at 576–77.
 Bruce G Peabody, “George Washington, Presidential Term Limits, and the Problem of Reluctant Political Leadership” (2001) 31:3 Presidential Studies Q 439 at 445.
 Stephen W Stathis, “The Twenty-Second Amendment: A Practical Remedy or Partisan Maneuver?” (1990) 7 Const Commentary 61 at 63, n 11, quoting Paul B Davis, Presidential Tenure: A Review and Reappraisal (PhD Dissertation, University of Utah, 1978) at 43.
 See Stathis, supra note 304 at 63.
 Tom Ginsburg, James Melton & Zachary Elkins, “On the Evasion of Executive Term Limits” (2011) 52:6 Wm & Mary L Rev 1807 at 1834.
 Jack M Beermann, “A Skeptical View of a Skeptical View of Presidential Term Limits” (2011) 43:4 Conn L Rev 1105 at 1109.
 See Julia R Azari & Jennifer K Smith, “Unwritten Rules: Informal Institutions in Established Democracies” (2012) 10:1 Perspectives on Politics 37 at 44.
 David A Crockett, “‘An Excess of Refinement’: Lame Duck Presidents in Constitutional and Historical Context” (2008) 38:4 Presidential Studies Q 707 at 710–11.
 One might suggest that the president’s decision to forego a third term amounts to a circumvention of the constitution’s requirement that an Electoral College select the president. See US Const art II, § 1, cl 2–3; US Const amend XII. On this view, the decision to abstain from a third term would circumvent the text by denying the Electoral College the free choice of whom to select as president. The argument would be more plausible where the president had decided to forego a third term in order to bind successors to the same choice.
 See Albert, “Constructive Unamendability”, supra note 10 at 211–15.
 See Constitution Act, 1982, supra note 6, ss 38, 42.
 See Albert, “Constructive Unamendability”, supra note 10 at 211–15.
 See ibid.
 See Part III.B, above.
 See Peter Suber, The Paradox of Self-Amendment: A Study of Logic, Law, Omnipotence, and Change (New York: Peter Lang, 1990) at 21.
 Richard Albert, “Nonconstitutional Amendments” (2009) 22:1 Can JL & Jur 5 at 5.
 Melissa Schwartzberg, Democracy and Legal Change (New York: Cambridge University Press, 2007) at 6.
 See Richard Albert, “Constitutional Handcuffs” (2010) 42:3 Ariz St LJ 663 at 665.
 See Richard H Fallon, Jr, “Legitimacy and the Constitution” (2005) 118:6 Harv L Rev 1787 at 1795.
 See ibid at 1794.
 There is a third form of legitimacy worth noting—moral legitimacy—which the change may possess if it is morally justifiable and worthy of respect. See ibid at 1796–97. But this is a content-based feature not assessable without more information on the given amendment. It is possible, in fact, for an amendment to possess both legal and sociological legitimacy without possessing moral legitimacy.
 See Elster, supra note 33 at 99–105.
 See Christopher L Eisgruber, Constitutional Self-Government (Cambridge, Mass: Harvard University Press, 2001) at 19.
 In his study of time and self-government, Jed Rubenfeld theorizes this temporal dimension to constitutional change. See Freedom and Time: A Theory of Constitutional Self-Government (New Haven: Yale University Press, 2001) at 11–12, 175–76. I have elsewhere argued that major constitutional reform should be attentive to the intergenerational binding quality of formal amendment. See Albert, “Amending Constitutional Amendment Rules”, supra note 37.
 One could argue in contrast that the views of political actors and the relevant publics would become internalized in the choice of the prime minister to continue nominating the consultative election winners to the Senate—and that the procedure should therefore be seen as supported by democratic choice. The problem with this position, however, is that the Constitution of Canada does not authorize a change to the method of senatorial selection in this way: it requires such a change to occur by formal amendment alone.
 See James Ross Hurley, Amending Canada’s Constitution: History, Processes, Problems and Prospects (Ottawa: Canada Communication Group, 1996) at 14–18.