Introduction

The recent actions of federal electoral boundary commissions and Parliament have together transformed Canada’s electoral map. In 2014, non-partisan, independent commissions established new boundaries for federal ridings. The new map was in place for the 2015 general election and replaced the one that had been in effect for the votes in 2004, 2006, 2008, and 2011. The commissions set the riding boundaries in accordance with the number of seats assigned to each province by the Fair Representation Act[1] of 2011. The Fair Representation Act increased the size of the House of Commons from 308 to 338 by creating 30 additional seats and distributing them to the most populous provinces, namely Alberta, British Columbia, Ontario, and Quebec.

Electoral districts are one of the “building blocks” of representative democracy.[2] Given the redistribution of seats among the provinces and the readjustment of electoral boundaries within them that has just been completed,[3] it is an appropriate juncture to reconsider how we design these building blocks of democracy.[4] There are two relevant dimensions here: (1) the manner in which district boundaries are set and (2) the distribution of seats in the federal House of Commons to the provinces. This article focuses on the first dimension and, in particular, on the use of the discretion held by electoral boundary commissions to establish districts within each province.[5]

The system of non-partisan, independent commissions was instituted in 1964, and the legal framework overseeing their decision making has remained largely unchanged ever since. One commission is established to set federal districts in each province, for a total of ten decision-making bodies operating under a common legal framework. Extensive freedom of action within the realm of electoral boundary design, with few statutory or constitutional constraints, characterizes the discretion held by the commissions. The introduction of commissions was designed to combat the entrenched practice of partisan gerrymandering that successive governments from across the political spectrum enthusiastically carried out from Confederation onward.[6] The institutional move away from Parliamentary control of redistricting was so significant as to earn the title of an “electoral boundary revolution.”[7] The electoral boundary revolution ended the direct partisan manipulation of electoral boundaries by elected representatives.[8]

The extensive discretion granted to commissions to set boundaries, however, generates a new series of potential problems that can undermine the fairness of the electoral map. This article analyzes the use of discretion by federal electoral boundary commissions and poses several questions: What type of discretion do boundary commissions possess? How have the commissions interpreted and applied this discretion? Has the ample leeway for discretionary decision making granted to the commissions furthered the fundamental values underlying Canadian redistricting? To address these questions, this article considers in particular the federal electoral maps set in 2004 and 2014, and the interim and final reports of the commissions operating at those times.

I argue that the commissions’ use of their discretion has fractured the right to vote. Malleable statutory rules, a vague constitutional framework related to electoral boundaries, and the subsequent uncertainty about how to translate democratic principles into practice have allowed commissions to adopt a multitude of distinct approaches to their common task. Many of these approaches are not complementary, but mutually incompatible. The right to vote is brought to life by the legal rules that govern the democratic process, such as those involving district design. The exercise of discretion by commissions in competing and contradictory ways has reached such a degree of variation that Canadians unevenly enjoy the right to vote.

The discretion granted to commissions should be restructured in order to ensure greater coherence and consistency in electoral maps. I do not advocate for the elimination or minimization of commission discretion, as such changes would frustrate the important values that underlie the use of administrative bodies to conduct activities formerly conducted by legislatures. Rather, their discretion should be reshaped to encourage the equality of voters, while also preserving flexibility for commissions to balance relevant factors in their decision making. Changes in the institutional makeup and procedures of commissions may also be necessary to further democratic values. The institutional reform agenda launched by the electoral boundary revolution in 1964 remains unfinished.[9] By restructuring the commissions’ discretion, the full promise of the reforms launched in 1964 can be fulfilled.

The article proceeds as follows. Part I considers what type of discretion electoral boundary commissions do in fact hold. I argue that the commissions formally possess what Ronald Dworkin labels “weak discretion”, but in practice exercise a greater degree of discretion than this category would initially indicate. Part II claims that a fracturing of the right to vote is problematic because it violates the principle of the equality of all voters and leads to arbitrary treatment. Part III considers in detail the competing approaches adopted by electoral boundary commissions to flesh out this claim about the fracturing of the right to vote. Part IV raises various reform options including amending the Electoral Boundaries Readjustment Act[10] to implement a hierarchy of criteria approach that has been used fruitfully for redistricting in some states in the United States and altering the structure of commissions.

I.   What Type of Discretion Do Electoral Boundary Commissions Possess?

A.  The Structure and Composition of Commissions

The creation of the boundary commission system was a direct reaction to the plague of gerrymandering that was endemic in Canadian federal elections. Partisan gerrymandering means the deliberate design of electoral boundaries in order to maximize the competitive advantage for one political party or candidate against others. The prevalence of partisan gerrymandering is a defining feature of redistricting in the United States.[11] While the dark art of gerrymandering may have reached its apotheosis there, the history of electoral boundaries in Canada prior to 1964 is similarly troubling. Before the move to commissions, redistricting was often a blatant exercise in self-dealing by Members of Parliament (MPs) seeking to gerrymander districts for either partisan gain or to protect incumbents. Massive deviations in riding populations were the norm and reforms that would equalize constituency populations were thwarted in order to create districts that favoured rural voters and politicians.[12] As John Courtney has pointed out, the nine federal redistributions and redistrictings between 1872 and 1952 were “carefully managed by the government of the day, whether Conservative or Liberal, in its own interest. The great majority … were partisan and blatantly self-serving affairs.”[13]

The EBRA of 1964 implemented the commission system. There had long been calls for a non-partisan process, but governments of the day consistently found that they had less appetite for such reforms once in power.[14] A series of minority governments created conditions where each of the two major parties, the Liberals and Conservatives, could see the benefits of ending gerrymandering.[15] The appropriate degree of partisan involvement in the drawing of boundaries, if any, was at issue from the beginning of the debate around the EBRA. An early version of the bill would have created four-member commissions composed of the Representation Commissioner (a non-partisan public servant), one political appointee from each of the Conservative and Liberal parties, and a judge nominated by the Chief Justice of the province. The Representation Commissioner was to sit on each of the ten provincial commissions in order to bring consistency and expertise to the process[16] and to draft the initial maps in each province.[17] The smaller opposition parties objected to being excluded and the bill was amended. Rather than partisan appointments directed by the parties, the task of nominating the two commissioners other than the Representation Commissioner and the judge was given to the Speaker of the House of Commons. The Chief Electoral Officer took over most of the powers of the Representation Commissioner in 1979, which led to the Commissioner’s position being eliminated. As a result, the original four member commissions were reduced to three, without any overlap in membership across the provincial commissions. This structure remains today, with the judge chosen by the Chief Justice of the province as chairperson and the other two members chosen by the Speaker. These two additional members have often been drawn from legal and academic ranks.[18] Boundary drawing had moved beyond an ends-driven, partisan-motivated activity to one where public-minded criteria could be applied and balanced.

B.  Types of Discretion

In this section, I consider the grant of discretion to the commissions by the EBRA, which allows them to set riding boundaries without having to gain approval from Parliament. How should we understand the idea of discretion, in order to determine what type is granted to electoral boundary commissions? In Taking Rights Seriously, Ronald Dworkin argued that discretion is engaged “when someone is in general charged with making decisions subject to standards set by a particular authority.”[19] Dworkin set out two types of discretion that may be granted to a decision-making body: strong discretion and weak discretion.[20] He acknowledged that these are “gross distinctions” and that “the shadings are many” in between the two categories.[21]

Strong discretion exists where the decision maker possessing delegated power “is simply not bound by standards set by the [authorizing] authority in question.”[22] Strong discretion is, in other words, the ability of a decision-making body to choose the rules that it will apply. This type of discretion can exist even where a higher authority will review the decisions of a body,[23] as is the case when courts oversee the actions of an administrative agency such as a boundary commission.

Weak discretion, in Dworkin’s formulation, involves the authority to make decisions within rules set by an entity higher up the decision-making hierarchy. Weak-form discretion still entails the use of judgment by the decision maker, but according to obligatory criteria set from above. There is discretion in the application of these criteria, but not in the choice of whether to apply them or whether to select alternative standards. Dworkin provides the examples of the sergeant ordered by a superior to pick her most experienced soldiers or the boxing referee obliged to declare the more aggressive combatant to have won the round as classic examples of weak-form discretion.[24] In both cases, the decision maker has leeway to arrive at an outcome, but must do so by recourse to an externally supplied benchmark.

In applying Dworkin’s typology, electoral boundary commissions can be said to formally possess weak discretion. The discretion exercised by the federal commissions exists within the bounds set primarily by the EBRA. Commissions are also subject to the constitutional rules on redistricting emanating from the right to vote. The EBRA empowers the commissions to balance a specific set of factors that direct to what ends their authority should be exercised. Stepping outside of the statute will constitute ultra vires activity by a commission. In being obliged to follow the specific factors set out in their home statute, the EBRA, Canadian boundary commissions are no different from their counterparts such as the Australian federal redistricting commissions, which must balance the factors established in the Commonwealth Electoral Act 1918,[25] or those in the United Kingdom, which apply the Parliamentary Constituencies Act 1986.[26] All can be said to formally possess weak-form discretion.

      In practice, however, the formal grant of weak discretion to Canadian federal commissions does not capture how they actually operate. The competing standards set out in the EBRA, nearly all of which are vague and abstract, provide the commissions with the freedom to act with something beyond the weak-form discretion than an initial glance at the legislation would indicate. The rules set by Parliament for the commissions to apply in many cases cannot be simultaneously maximized and are often in conflict. The EBRA provides ample space for commissions to select factors that they wish to emphasize and to diminish the importance of others. As a consequence, the decisions by the ten commissions in applying the EBRA are so disparate, and at times opposing, that it often appears they are not really working from the same set of rules. Electoral boundary commissions in practice possess something in between weak and strong discretion. They exist as one of Dworkin’s “shadings” between the two categories.

The EBRA addresses the two major dimensions of commission discretion—how to set the populations of ridings and how to take into account communities of interest, including minority representation. The EBRA requires commissions to design districts “as close as reasonably possible”[27] to population equality, which reflects a commitment to representation by population. When districts have unequal populations, individuals in the most populous ones will have their voting power diluted. Creating districts of relatively similar populations is therefore a reflection of the principle of the equality of all citizens.

The commitment to representation by population reflected in the legislation is tenuous, however, as the EBRA also sets out grounds for deviating from it. The EBRA directs that the commissions “shall consider … community of interest or community of identity,” the “historical pattern of an electoral district,” and the need to create districts of “manageable geographic size.”[28] These are explicitly established as the only grounds for departing from representation by population.[29] Deviations from voter parity must be within a range of twenty-five per cent above or below the average riding size in the province, though they can be even greater in “extraordinary circumstances.”[30] The permitted variance of twenty-five per cent or more in extraordinary circumstances is an outlier when put in the context of other comparable democracies[31] and some Canadian provinces,[32] which provide less scope to deviate from representation by population.

A commission is somewhat like the sergeant commanded by a superior to pick her most experienced soldiers for a task. They are given discretion within a limited scope. Yet the commissions depart from the weak-form model, because the variety of factors that they may consider enables them to pick and choose which to apply. Commissions sometimes prioritize representation by population over community of interest, while others reverse that hierarchy of decision-making criteria. Commissions are therefore better described as a sergeant commanded to select the best soldiers for a mission, taking into account experience, recent performance regardless of experience, training, bravery, and other factors, without guidance as to how to balance the criteria set out by the individual higher up in the hierarchy. Predictably, if multiple officers are provided this set of decision-making parameters, different approaches to the weight appropriately given to each criterion will emerge. Some will prioritize experience or recent performance above the other factors. Some sergeants may put equal weight on all criteria.

Electoral boundary commissions are in the same situation. The requirement to balance various criteria, many of which are subject to disputes about their definition and meaning, amounts in practice to a minimal constraint on their discretion. While commissions certainly require leeway to determine where best to place boundary lines given local conditions, they in effect wield power to determine what principles should govern.

The effects of the statutory scheme are compounded by the Supreme Court of Canada’s interpretation of the right to vote in section 3 of the Charter.[33] Like the statutory scheme, the constitutional rules surrounding redistricting facilitate something beyond weak-form discretion. The Court has interpreted section 3 to guarantee “effective representation” in the context of drawing electoral maps.[34] Outside of electoral boundaries, the Court has applied a meaningful participation standard.[35] The most directly applicable authority remains Reference Re Provincial Electoral Boundaries (Sask)[36] (also known as Carter), which is the only redistricting case to reach the Court in the Charter era. The Court has never ruled directly on the constitutionality of the EBRA.

At issue in Carter was a Saskatchewan provincial electoral map that displayed large deviations from representation by population to the benefit of rural voters and at the expense of urban ones.[37] The majority found that effective representation rather than representation by population was the proper meaning of section 3.[38] Effective representation was understood by the majority to entail voter parity as the primary concern, but to permit deviations for reasons such as “geography, community history, community interests and minority representation” in order to “effectively represent the diversity of our social mosaic.”[39] The recognition of voter parity as the first principle, even if it rejected strict representation by population, implied that the majority would contemplate some levels of variance as too great. In other words, not any riding population would be constitutionally acceptable.

The majority, however, did not provide a number as to what level of deviation would fall within the parameters congruent with the guarantee of effective representation. The Saskatchewan map did involve significant deviations,[40] but the majority refrained from setting any clear guidelines as to whether similar levels would always be acceptable. The goal of the justices in the majority appears to have been to supply a template to guide courts in future challenges to electoral boundaries, which necessarily had to be broad and abstract so as not to limit its general application.[41] Whatever the merits of permitting deviations from representation by population,[42] or lack thereof,[43] moving away from a one person, one vote bright line constitutional rule had the secondary effect of expanding the room within which commissions had to act without fear of running afoul of the Charter.

Because of the failure to impose anything other than vague standards either under the EBRA or under section 3 of the Charter, commissions may act largely unconstrained in the exercise of their discretion. Commissions must be cognizant that truly extreme departures from voter equality are unconstitutional. They must be aware that they are required to justify deviations with reference to one of the relevant factors listed in the statute or highlighted by the Court in Carter. Beyond these parameters, however, commissions retain extensive freedom of action. Neither the statutory scheme nor the constitutional rules encourage weak-form discretion in practice.

II. Political Equality

In Part III, I will demonstrate the specifics of how commissions have interpreted their expansive discretion in widely varying ways. Before turning to that argument, this Part will first address the principle of political equality that emerges from the Supreme Court of Canada’s jurisprudence on democratic rights and freedoms. That federal boundary commissions adopt conflicting approaches is troubling because it fractures any common realization of the right to vote and thereby offends this principle. Given that Carter is the only redistricting case from the Supreme Court in the Charter era, the principle of political equality is largely embedded in the case law on areas other than electoral boundaries. The Court’s incorporation of the idea of political equality crosses multiple areas of the law of democracy and includes section 3 as well as section 2(b).

The notion of the equality of voters is central to section 3, which guarantees to “[e]very citizen” the right to vote and to stand for elected office in federal and provincial elections. In the jurisprudence on the scope of the right to cast a ballot under section 3, the courts have reiterated that there are no internal limits to this broad and inclusive language, beyond the explicit requirement of citizenship, in order to qualify for the right.[44] No distinctions on the basis of “privilege or merit” are permitted, as the right to vote attaches automatically to all citizens as “a function of membership in the Canadian polity that cannot lightly be cast aside.”[45] Bans on voting by prisoners,[46] those of diminished mental capacity,[47] and judges[48] have therefore been struck down in the Charter era.

The Court’s campaign finance jurisprudence engages with a different context than redistricting, but equality is a central concern. The Court has acknowledged that limits on spending by participants during a referendum[49] or by third parties during the election period[50] can be justified under the “egalitarian model”.[51] The logic behind the egalitarian model is that unlimited spending would allow those with resources to dominate the election campaign. The egalitarian model endorsed by the Court seeks, instead, a level playing field so that all voters will be on similar footing. Some overly onerous limits on political spending will offend the freedom of political speech guaranteed by section 2(b) of the Charter, as was the case in Libman v. Quebec,[52] but the Court has undoubtedly emphasized the equality of voters above other considerations in campaign finance.

In R v. Bryan,[53] the Court extended the egalitarian model so far as to guarantee “informational equality” to voters. At issue in the case was the constitutionality of section 329 of the Canada Elections Act[54] limiting the premature transmission of election results from Eastern Canada before the polls had closed in the West. The Court upheld the conviction of Bryan for violating section 329 on the grounds that the provision guaranteed informational equality to voters in British Columbia. It would be unfair and harmful to democratic participation, the Court reasoned, to furnish voters in British Columbia with information about election results that might shape their behaviour, on whether to vote and for whom, that had not been available to those in the East.[55]

Equal treatment has also permeated the Court’s jurisprudence on political parties. In Figueroa, the Court struck down rules discriminating against small political parties fielding less than fifty candidates.[56] The majority for the Court reasoned that parties are the collective vehicle that facilitates meaningful individual participation. The case can be understood as standing for the principle of equal treatment of political parties, regardless of size, popular support, contribution to political debates, or electoral success.[57]

      Carter itself is a qualified exception to the emphasis on equality running through the rest of the jurisprudence on democratic rights and freedoms. The majority in Carter rejected the idea that all individual votes should be worth the same, but can be said to have endorsed the equal treatment of similarly situated voters. Deviations from representation by population were permitted in the case. The Court did not, however, exclusively grant this advantage to rural voters. The Court included any group falling within the “social mosaic”[58] as potentially deserving of the same treatment if it would lead to the more effective representation of the electorate as a whole. This reasoning did harm to the principle of representation by population, but the Court’s theory of representation cannot be said to ignore the principle of equality entirely. All groups of citizens forming part of the social mosaic could potentially benefit from the doctrine of effective representation. There were in fact calls after Carter, taking up this line of reasoning, to deviate from representation by population to aid minority groups other than rural voters.[59]

The commissions’ contradictory interpretations of the legal framework result in concrete discrepancies. The principles governing redistricting determine (1) voting power and (2) how individuals are aggregated into political communities in the form of ridings. When one commission permits the underrepresentation of urban voters in a province, but another ensures representation by population for the same group, then the voting power of city dwellers will vary according to the whims of the decision-making body. Carter certainly envisioned a fairly wide latitude for deviations from representation by population. It did not endorse, however, a view that urban voters might have radically different voting power depending on the province in which they live.

The same goes for the aggregation function of electoral map making. It may be that, on the ground, local conditions dictate that municipal boundaries, racial identity, or a common economic status are relevant factors to take into account in assessing what communities to group together into ridings. Having one commission reject outright that some of these factors are relevant, while another insists that they must be of primary consideration, results in unequal treatment. Racial minorities may have their communities of interest recognized by a commission in one province but not another. The unequal treatment that results from the competing approaches adopted by commissions is reflected in diverging voting power and erratic recognition of communities of interest. These consequences are a constitutional problem, as contradictory approaches to electoral district design erase any consistent exercise of the right to vote.

Departures from political equality arise both within and across provinces. A commission in British Columbia may differ in its approach from one in Newfoundland. That raises problems of interprovincial equality. There may also be violations of the equality principle within provinces. A commission will sometimes adopt a dramatically different interpretation of the relevant principles of redistricting than its predecessor from ten years earlier, despite the EBRA, the constitutional framework, and the demographic features of the province remaining constant.

One could object that the principle of federalism should infuse boundary design in Canada and, therefore, that competing approaches by commissions across provinces are to be encouraged to foster respect for local and regional differences. The existence of one commission per province would seem to supply some evidence of a federal intent behind the EBRA. Yet the EBRA deliberately implemented a uniform legal framework for all commissions, a common structure, and set procedures regarding consultation, the production of reports, and timelines. The introduction of the Charter in 1982 provided a further set of common constitutional principles applying to all commissions. The Canadian approach of a common set of rules differs from that in the United States, where federal law and constitutional provisions apply, but each state controls the process and can introduce its own rules as long as they do not displace the federal ones. Canada’s contrasting constitutional design choice suggests coherence rather than fragmentation was the goal.

This approach reflects an important underlying democratic principle. All representatives are elected to serve in the same House of Commons. One way of viewing the competing approaches to designing districts is that representatives from one province are elected on slightly different bases from those in another, if the commissions endorse divergent principles. Different bases of representation create a more subtle distinction than if each province were to elect representatives to the House on the basis of alternative electoral systems, with Ontario using first-past-the-post and Quebec proportional representation, for example. Yet it does produce differences in how representatives are elected that goes beyond simply accounting for local conditions. Having representatives within the same legislative body being elected according to different rules violates the equality of voters under federal electoral law and section 3 of the Charter, and the equality of representatives who formally hold the same status in the House.[60]

Allowing contradictory approaches to redistricting in order to further the principle of federalism also misunderstands the relationship between redistricting and boundary redistribution. Federal concerns are accommodated through the rules governing how many seats each province is entitled to in the House. The addition of thirty new seats through the Fair Representation Act reflected a change in the formula for how seats are apportioned. Constitutional rules exist to protect the entitlements of the less populous provinces, with Prince Edward Island, for example, maintaining four representatives regardless of its population.[61] Federalism is respected through the rules setting out the interprovincial distribution of seats. The federal principle, however, should not trump political equality in the design of districts within each province.

III.               Conflicting Approaches to Commission Discretion

In light of the emphasis on the principle of political equality in the jurisprudence on democratic rights and freedoms, the exercise by commissions of expansive freedom of action closer to strong- than weak-form discretion is problematic. The stronger the discretion held by boundary commissions, the more they are able to determine which principles to follow in setting electoral districts. In combination with the decentralized model, where one commission exists for each province for federal districts, this type of discretion facilitates a proliferation of different approaches to how to draw boundaries. These approaches vary so widely that they can be said to differ not only in their interpretation of the relevant principles, but in the selection of which principles to apply. The commissions have adopted not just different methods, but competing and contradictory ones.

The impact of expansive commission discretion, as a result of the EBRA and Carter, is that the right to vote, as filtered through electoral districts, is enjoyed unevenly across the country. Similarly situated voters with equal claims to fair representation see their treatment vary widely despite a common statutory regime and constitutional right. No commission entirely rejects principles such as representation by population or community of interest in the abstract. But radically different approaches currently coexist as to how to interpret the factors designed to guide redistricting. These contradictory approaches are not confined to trivial matters, but affect issues that go to the heart of democratic representation in an electoral system of geographic districts. The result is that the right to vote is stretched beyond any coherent meaning. Particularly when understood in the context of the Court’s jurisprudence emphasizing the equality of voters, the right to vote cannot be seen as so elastic as to encompass dramatically different results for similarly situated voters.

Some minor variation in how the right to vote is interpreted is inevitable between commissions facing particular circumstances on the ground. There may be relevant distinctions between voters or local conditions that commissions should take into account. If a community has a predominant economic occupation, or if voters with other common interests cluster together, then commissions may reach different results than in the absence of these local conditions. I am not making an argument for viewing Canadian voters as undifferentiated citizens shorn of other identities. If a certain ethnographic trait is a relevant factor to take into account in redistricting, however, then it should hold as much in one part of the country as another. If the EBRA stated that race was a relevant factor to consider in drawing the map in Manitoba, but not British Columbia, then it would create inequalities. Just as troubling would be a rule that linguistic minorities should have their electoral influence maximized in Nova Scotia, but not in Quebec. Strict enforcement of representation by population in British Columbia should not coexist with manifest disregard for that principle in Ontario, whatever the merits of either approach. Such variations, however, occur in practice.

In order to flesh out my claim about the fracturing of the right to vote, I turn now to outlining the variety of interpretations that have emerged. I trace these conflicts, and their impacts, on the two main areas of commission discretion—representation by population and community of interest concerns.

A.  Representation by Population

Requiring boundary drawers to stick as closely as possible to representation by population, but then providing an expansive range of variance and a list of exceptions, enables commissions to differ in their interpretation of the acceptable range of constituency populations. The spectrum among recent commissions spans from strict adherence to representation by population to a competing view of representation by population as a secondary concern.

1.   Strict Adherence to Representation by Population

Some commissions adhere to representation by population in a fashion that brings Canadian boundaries closer to international norms. These commissions take seriously the obligation in the EBRA to adhere “as close as reasonably possible” to creating districts of equal population and the statement in Carter by the Supreme Court that voter parity is the primary factor to consider. The logic adopted by these commissions is that representation by population, as the dominant principle in drawing boundaries, should be departed from as infrequently and as lightly as possible. Three distinct tools are adopted by commissions seeking to uphold representation by population: (1) minimizing deviations in the map as a whole; (2) rejecting the use of the extraordinary circumstances clause; and (3) pre-emptively establishing a permissible range of deviation well below that permitted in the EBRA.

The Alberta commission from the 2014 redistribution is a clear example. Despite the province’s remote northern region and a large territory, all but one of Alberta’s electoral districts in the initial map proposed by the commission was within a five per cent threshold. The sole exception was the district of Fort McMurray-Cold Lake, which had a population 5.29 per cent below the provincial average.[62] A riding with a deviation of 5.29 per cent would be as close as some commissions come to representation by population, while for the 2014 Alberta commission it was the maximum permitted.[63] After the public consultations, the Alberta commission agreed to two additional ridings over the five per cent threshold,[64] but these were only marginally outside of the range and all ridings remained within ten per cent of the average.

Another example of strict adherence comes from the evolving habit of disuse of the “extraordinary circumstances” clause. Commissions possess discretion about which districts the “extraordinary circumstances” clause can legitimately be applied to and even about whether to use the provision at all. Its use has dropped markedly in recent redistributions.[65] Only two ridings, Kenora-Rainy River and Labrador, exceeded the twenty-five per cent variance in the 2004 redistribution, and both had very strong claims to do so given their remote geography. The trend is for commissions to abstain from invoking the extraordinary circumstances exception.

In 2014, the New Brunswick commission displayed clear reasoning about why the extraordinary circumstances rule should not be applied, even if technically permitted by the statute.[66] The commission considered the extraordinary circumstances rule to be incompatible with its mandate from the EBRA in all but the rarest of cases, where ridings are of truly unmanageable geographic size. It refused in advance to apply the provision on the logic that New Brunswick’s relatively compact geography could not constitute an extraordinary circumstance. The rejection of the extraordinary circumstances clause usually indicates that a commission has adopted this first model of strict adherence to representation by population.

Commissions in recent redistributions have also established the primacy of voter parity by explicitly setting lower variance thresholds than the twenty-five per cent range permitted in the statute. These commissions commit themselves in advance to permitting no deviations beyond five or ten per cent and then go about drawing the map. The New Brunswick federal commission in 2004 attracted the most controversy for taking this approach.[67] Its reasoning was that as representation by population is the primary criterion, deviations beyond ten per cent would be inappropriate in New Brunswick even if the EBRA granted them the power to do so. Other provinces have followed suit in setting lower variance targets.[68]

This upfront rejection of the statutorily permitted variance has been controversial and is worthy of further consideration. The New Brunswick commission’s exercise of its discretion was challenged in Raîche v. Canada (AG).[69] Raîche directly raises the question of whether establishing a lower variance target than the maximum permitted in the EBRA is permissible. At issue in Raîche was the boundary between two neighbouring northern districts, Miramichi and Acadie-Bathurst. Acadie-Bathurst was an eighty-five per cent majority francophone district because of its large Acadian population.[70] Miramichi was a majority anglophone district, but with an influential francophone minority of thirty-three per cent. The commission’s initial map moved some francophone voters from the majority-minority district of Acadie-Bathurst to Miramichi, where French-speakers were in the minority, in order to move closer to population equality and the ten per cent target.[71] Without this transfer between the ridings, Acadie-Bathurst would have been fourteen per cent above the provincial quotient and Miramichi would have been twenty-one per cent below. Although deviations of fourteen and twenty-one per cent were within the statutorily permitted twenty-five per cent range, they amounted nonetheless to a large variance of thirty-five per cent of the quotient between neighbouring, geographically compact districts.[72]

Opponents of the revised map claimed that by moving francophone voters into Miramichi, the commission failed to comply with its statutory obligations under the EBRA to respect communities of interest.[73] The claim rested on an interpretation of the EBRA that deviations from voter parity were required to respect the Acadian community of interest. The claimants asserted that as long as the deviations were within the twenty-five per cent range set out in the statute, the commission was obliged to depart from its stated goal of no more than a ten per cent deviation to ensure minority representation. The commission had reasoned that a ten per cent deviation target was a fair compromise between voter equality and the other factors listed in the EBRA. On the particular case of the boundary shift of Acadie-Bathurst, the commission explained that it considered the francophone minority in Miramichi to be sufficiently large that its voice would be heard and its interests represented.

Justice Shore found that the commission acted unreasonably. He concluded that there was no evidence the minority would be adequately represented in the Miramichi riding. His reasoning put great emphasis on the fact that the representative for Miramichi was a unilingual anglophone, while the representative for Acadie-Bathurst was a bilingual francophone. The commission’s willingness to adapt its original proposals by keeping within Acadie-Bathurst some of the francophone voters slated to be moved was held to be insufficient to discharge its statutory obligation. Fair representation for linguistic minorities is undoubtedly an important goal, especially in officially bilingual New Brunswick. Beyond the particular circumstances of the Acadian minority, a potentially negative consequence of Justice Shore’s reasoning, however, is that courts may feel there is precedent to restrict attempts by commissions to ensure strict adherence to representation by population in other scenarios.

The application of a ten per cent threshold raises the issue of whether the commission was unduly fettering its discretion as an administrative body.[74] Rigid, self-imposed rules curbing an administrative body’s own discretion may constitute unlawful fetters. Administrative bodies may impose limits on their own exercise of authority only as long “as [they do] not preclude the possibility that the decision maker may deviate from normal practice in light of particular facts.”[75] For the ten per cent rule to be lawful, it must not have prevented the commission from acting on relevant facts that would lead it to exceed the self-imposed threshold.

A careful analysis shows that the New Brunswick commission did not rigidly impose a ten per cent rule. In the face of criticism, the commission reversed some of its proposed changes. Its revised map accepted a greater than ten per cent deviation from the provincial quotient in the two ridings as well as the three large urban ridings in the province (Moncton-Riverview-Dieppe, Saint John, and Fredericton). The new maximum deviation was fourteen per cent. The final map did not reflect all the changes sought by Acadian groups, but it considered and responded to their concerns while seeking to strike a balance between community of interest and the principle of representation by population.[76] The changes from the draft to the final map demonstrate the New Brunswick commission used a ten per cent variance as a guideline, rather than a rigid rule, and that when presented with relevant facts it was willing to deviate from its own standard. It remains to be seen what degree of flexibility courts will require of boundary commissions in order to comply with the prohibition against self-imposed fettering of discretion.

2.   Representation by Population as a Secondary Concern

The strict adherence model coexists with commissions that can fairly be said to treat representation by population as a secondary matter. Some commissions have rejected the importance of representation by population, viewed any deviation within the twenty-five per cent range as acceptable, or been much more permissive in their use of the extraordinary circumstances clause. The 2014 Ontario commission stands out as a clear recent example.[77] The commission subordinated representation by population to other concerns, some of which were anchored in the exceptions found in the EBRA, while others had a more questionable legal basis. In comparison, in particular, to its counterparts in Alberta and British Columbia, which also faced the need to balance growing urban populations with sparsely populated rural and remote regions, the Ontario commission showed much less respect for the primary principle of redistricting.

That representation by population was treated as a secondary concern by the Ontario commission manifested itself across multiple dimensions. The commission prioritized artificially created regions over voter equality. It divided the province up into eleven regions,[78] many of which had sub-regions.[79] Although it is not clear how the commission decided what formed the boundaries of each region, its method was to assess the population of each region and then to assign it the appropriate number of ridings given the average population per riding that it was supposed to achieve.

Redistricting Ontario’s 121 ridings can be a daunting task. Changes to boundaries at one end of the province can have a ripple effect across it. Dividing the province into regions can be a useful heuristic device for map-makers seeking to render the task manageable, especially given the introduction of fifteen new seats, which obliged the commission to engage in wholesale changes.

Despite the potential usefulness of grouping ridings into regions, the commission unduly elevated the goal of fairness across regions above actual representation by population. With the exception of Northern Ontario, the commission granted each region something close to the allotment called for by representation by population. It was a different story, however, within each region. The regions of the “City of Toronto” or “Halton, Hamilton and Niagara” received the appropriate overall complement of seats, but the commission was very lax with regard to deviations between ridings within each of them. Fairness between regions trumped representation by population, as the commission refused to equalize riding populations across regional boundaries. Neither “regions” nor the principle of regional fairness are listed in the EBRA. The commission created the regions itself and provided little justification for their boundaries, but then used them as a justification for why following representation by population was impractical.

The consequences of this elevation of regional fairness above representation by population are stark. Niagara Falls and Niagara West are adjacent ridings in the “Hamilton and Niagara” sub-region of the “Halton, Hamilton and Niagara” region. Yet the Niagara Falls riding has a population of 128,357, or 20.85 per cent over the provincial average, while Niagara West has only 86,533 inhabitants[80]—18.53 per cent below the average. Voters in Niagara Falls are significantly underrepresented while those in Niagara West are overrepresented. The difference between the two ridings is nearly 40 per cent of the provincial quotient (or average riding population) at over 41,000 people. The concept of community of interest could potentially justify this change, in the abstract. It is fair to say, however, that populations across these two ridings are mobile. There is no mountain range or body of water separating them. There is no obvious difference in the demographic makeup of the two riding populations. The legitimacy of any community of interest claim that would justify the distinction between the two ridings is non-existent.

The Ontario commission also devalued representation by population in the largest urban areas, which it treated as distinct sub-regions, such as Ottawa and Toronto. Ottawa South has a population of 121,894, or 14.76 per cent above the average, while Rideau-Carleton’s population is only 89,522, or 15.71 per cent below,[81] for a spread of over 30 per cent. In the Greater Toronto Area, Burlington is underrepresented with 120,569 people, or 13.52 per cent above the quotient, while Don Valley East has a population of 93,007, or 12.43 per cent below,[82] for a 26 per cent range. Urban areas as a whole are underrepresented in Canada,[83] particularly in Ontario,[84] and the commission’s approach compounded that fact within the province’s largest cities. Deviations from representation by population have traditionally been justified because of the presence of remote, rural, or isolated populations. The EBRA’s requirement to create ridings of “manageable geographic size” provides support for some leeway for areas such as Northern Ontario. Without similar geographic challenges, there is no justification for such deviations within Ottawa and the Greater Toronto Area. The commission’s approach of prioritizing fairness across regions, and then assuming that representation by population required nothing further, led to mystifying inequalities of voting power within cities.

The secondary status of representation by population was further entrenched by the commission’s insistence that federal electoral districts should not cross municipal boundaries. As with regional lines, the EBRA again does not list municipal boundaries among the relevant criteria. It is possible to consider municipal boundaries as relevant in delineating communities of interest, due to shared local government. There is nothing wrong, in principle, with electoral districts at one level of government being tied to those at another. As with regional boundaries, however, the Ontario commission used municipal boundaries not just to help shape where to draw district lines, but to justify deviations from representation by population. If representation by population conflicts with the goal of incorporating municipal districts into federal ones, voter equality must triumph as it is the primary principle according to the EBRA and Carter. The Ontario commission made voter equality a secondary concern and elevated the preservation of municipal boundaries within federal ridings above it.

The emphasis on municipal boundaries meant that mid-size cities such as Kingston and Peterborough were among the most disadvantaged in the 2014 map, where they had not been under earlier ones. The commission created a whole new problem that had previously not existed, because it refused to cross municipal lines to combine populations residing in the official boundaries of these cities with those living just outside them. The height of this problem was in the riding of Brant, which was severely underrepresented with a population 24.70 per cent above the average, just below the extraordinary circumstances cut-off of 25 per cent.[85]

The commission did not try to hide its preferences despite the lack of authority for subordinating representation by population to them. It stated in its initial report that “[t]he Commission endeavoured to respect existing municipal boundaries whenever possible” and that “[i]t is virtually impossible to establish an electoral map for 121 electoral districts of equal population that reflects existing municipal boundaries.”[86] The commission failed to appreciate that the EBRA obliged it to drop its fascination with municipal boundaries.

B.  Community of Interest

The second major fault line for discretionary decision making, community of interest considerations, follows representation by population in being the subject of competing approaches by commissions. The result is a further fracturing of the right to vote. The community of interest principle is particularly controversial as it frequently engages issues of minority representation. How to incorporate minority communities within geographic districts with a single member plurality electoral system is a perpetual matter of debate in political theory.[87]

Commissions have adopted several competing ways of interpreting and applying community of interest concerns. The division here is over whether racial, ethnic, and linguistic minorities and Indigenous peoples constitute communities of interest. There are four distinct approaches evident from recent redistributions: (1) the undifferentiated citizenship model; (2) the creation of minority-influence districts; (3) affirmative gerrymandering; and (4) the design of non-contiguous districts to maximize minority influence. The term “community of interest” is not defined in the EBRA, which facilitates this proliferation of approaches.[88]

These four conflicting approaches emerge from the lacuna not only in the EBRA, which provides no clues as to the meaning of community of interest, but also in the jurisprudence. Carter validated the need to consider community of interest in designing electoral districts, as it listed that factor as one possible reason that may justify deviations from voter parity. There is a lack of clarity in the decision, however, as to what groups of voters qualify as having a community of interest and what the implications are of finding that a community of interest exists. The Court only states that boundaries should reflect “the diversity of our social mosaic.”[89]

In the 1990s, there was an extensive public debate about the meaning of community of interest. The Royal Commission on Electoral Reform and Party Financing[90] (the Lortie Commission) defined community of interest to include traditional concepts, including local boundaries such as school boards, geographic factors like the presence of mountains or rivers, and the history of a district. The Lortie Commission also included “demographic and sociological” characteristics as informing community of interest.[91] “Demographic and sociological” traits meant ethnic, racial, religious, and linguistic minorities could be communities of interest and that electoral boundaries could be drawn so as to improve their representation. In the wake of the Lortie Commission and Carter, discussions abounded about guaranteed representation for underrepresented groups.[92]

The House of Commons Standing Committee on Procedure and House Affairs subsequently considered amending the EBRA in the early 1990s to include a specific definition for community of interest.[93] The Reform Party proposed a definition that excluded demographic and sociological factors entirely. The bill that was eventually presented in the House, C-69, was more expansive than the Reform proposal as it included urban, rural, and Indigenous identity in the definition, but it specifically excluded any mention of race, religion, ethnicity, or language.[94] Bill C-69 died on the order paper when an election was called. Boundary commissioners were left with no greater clarity as to the meaning to ascribe to the statutory term “community of interest” than had been endorsed by the Court in Carter. As a result, commissions retain broad discretion in interpreting the term.

1.   The Undifferentiated Citizenship Model

Some commissions ignore minority representation entirely and do not consider these groups to form communities of interest. They consider traditional criteria to be relevant, but not socio-demographic traits such as race, ethnicity, or Indigenous status. Traditional factors include the reach of local media, such as newspapers, and geographic features, such as lakes and mountains. These commissions view individuals as undifferentiated for the purposes of drawing boundaries. Although a rationale is not always provided, the underlying claim is that group identity is irrelevant to representation.

The most explicit defence of this position comes from the Alberta commission in the 2004 redistribution. During the consultation phase, members of the public made submissions that the working class and immigrants formed communities of interest that should be recognized by the commissions in their map making. The commission responded that it disagreed “philosophically” with using “ethnicity or class” as a factor in drawing boundaries.[95] The commission stated bluntly that it “rejects the narrow conception of identity politics as a basis for drawing boundaries.”[96] Its underlying assumption seemed to be that all individuals are unhyphenated equals and that considering factors such as race is divisive.[97]

Analysis of the Alberta map demonstrates that the commission did not live up to its own philosophical statement, or applied it selectively to recognize some common identities and not others. The commission’s report dismisses claims by ethnic or class-based communities, but then recognizes agricultural, rural, urban, and suburban voters as distinct groups forming communities of interest. The Alberta commission recognized communities based on location (rural, urban, suburban) or economic activity (agricultural), but not the interests of minority ethnic voters or other economic markers such as class.

Another example where commissions have rejected minority representation as relevant to community of interest has occurred with respect to Indigenous peoples. Commissions have unfortunately often failed to take into account the rights of Indigenous peoples. They have, for example, often split communities into multiple ridings.[98]

The evolution of Saskatchewan’s federal boundaries provides a further illustration, though involving a denial that urban voters, rather than a minority, formed a community of interest.[99] Successive commissions in Saskatchewan had chosen to create mixed urban-rural districts in the major cities of Saskatoon and Regina. Saskatchewan is largely rural outside of these two urban hubs. Urban voters in these cities were placed in ridings combining them with the population in outlying, rural areas. The province has been alone at the federal level in purposefully designing mixed urban-rural districts. The broad implication was that urban votes were diluted. The Saskatchewan boundaries discriminated against the urban community of interest to such an extent as to affect the outcomes of elections. The dilution of urban votes had clear partisan implications given party politics in the province.[100] The 2014 commission finally ended this practice of mixed urban-rural seats by recognizing an urban community of interest distinct from the rural one.

2.   Minority Influence

Other commissions have sought to enhance minority representation by drawing lines that keep groups of voters together so as to maximize their influence within districts of relatively equal population. These commissions create what can be described as minority influence districts, where a significant enough percentage of the electorate is made up of minorities to shape electoral results. For example, the Ontario commission in the 2004 redistribution responded to presentations by local communities at the consultation stage by amending their original map to keep the Portuguese community in the west-end Toronto riding of Davenport intact, rather than split across ridings.[101] The boundary lines were justified on a community of interest basis.

Often this approach occurs below the surface. It is a long-standing practice that commissions in diverse provinces with large numbers of racial minority voters tend to take the presence of such groups into account. District populations are kept relatively close to one another, and there may be no grand philosophical statements regarding theories of minority representation, but boundaries are often drawn to keep minority communities together.[102]

3.   Affirmative Gerrymandering

Commissions have also gone beyond creating minority influence districts to engage in what has been called affirmative gerrymandering. This approach attempts to maximize the ability of minority voters to select representatives of their choice. It does so by creating districts with a majority from a group that is a minority in society or by keeping individuals with a shared trait together even if that means deviating from representation by population. The most prominent example comes from the Nova Scotia provincial commission of the 1990s. The Nova Scotia commission engaged explicitly in something very close to affirmative gerrymandering of the type required for African Americans and Latinos under the American Voting Rights Act.[103] The VRA requires congressional districts that maximize the ability of minority voters to elect representatives of their choice. In practice, it requires a number of majority-minority districts consistent with this goal given the population and concentration of minorities in each state.

The Nova Scotia commission deviated significantly from representation by population to create districts where Blacks and Acadians had significant influence.[104] The commission justified this approach based on the need to ameliorate the chronic underrepresentation of these minorities. The Nova Scotia experiment stands out as the clearest example of affirmative gerrymandering among recent commissions. The goal was explicitly to improve the lot of minorities, despite the constraints imposed by geographic districts, by valuing the election of representatives chosen by substantial numbers of voters from these groups above representation by population. The claim by the Nova Scotia commission was not that representation by population was irrelevant or of secondary concern generally, but that in the face of persistent and pernicious underrepresentation of minorities, affirmative gerrymandering was an appropriate response.[105] The commission put weight in its deliberations on the traditional underrepresentation of these communities, their lack of influence with the government, and their mistreatment by the political majority.

4.   Minority Representation Through Non-Contiguous Districts

A small strand of commission reasoning has even attempted to decouple representation of minority communities of interest from any geographic constraints. It is infrequent in Canada, but it does exist. The 2004 New Brunswick commission proposed creating a non-contiguous riding incorporating all of the province’s Indigenous population, mainly composed of the Mi’kmaq and Maliseet. The idea was not supported by Indigenous peoples in the province and was therefore appropriately dropped. It is one type of response where a group is not geographically concentrated enough to benefit from the creation of minority-influence or majority-minority districts, or affirmative gerrymandering.

IV. Reforming Commission Discretion

The preceding analysis implies that boundary commissions have exercised their discretion in conflicting ways and that reforms are required to further the equality of voters. In this Part, I discuss options for reshaping discretion and for amending the commission structure in a manner that would bring greater coherence. I consider changes to the EBRA’s redistricting criteria and to the composition of the commissions.

The claim that reform is necessary runs counter to some leading work on electoral boundaries in Canada. Ron Levy has advanced the most strenuous defence in the literature of expansive discretion in the hands of redistricting commissions.[106] Levy argues that legal analysis tends to favour command and control structures that oblige entities to act consistently with some agreed upon norm. This constraint model, as he calls it, is based on the assumption that limiting discretion furthers the rule of law.[107] He identifies the assumptions of the constraint model as being that (1) administrators overseeing election law will amass power for themselves or their allies and (2) that their authority must be severely limited in consequence.[108] In his view, the constraint model overestimates the fear of abuse of discretion.[109] The granting of broad discretion to the electoral boundary commissions runs counter to the instincts of the constraint model. Levy argues this discretion is beneficial as it operates as a bulwark against manipulation of the democratic process by elected representatives. Because there are few precise, bright line rules governing redistricting in Canada, commissions operate relatively free of interference by partisans. Ambiguity surrounding the legal obligation to balance the competing factors listed in the EBRA is a virtue for Levy, rather than a failing, because it generates trust and encourages non-partisan decision making. In his formulation, the constraint model also places process over substance, thereby simply transferring the fight about outcomes to disagreements about procedures that dictate results.[110]

I am sympathetic to Levy’s general appreciation that we must guard against partisans using the redistricting process for political gains.[111] Levy is also correct that eliminating commission discretion in favour of rigid rules is neither feasible nor desirable. Absolute rules are likely only to mask the inevitable range of discretionary decision making facing commissions and would put at risk their ability to actually respond to local conditions and shifting demographics. The fear of discretion existing in a constraint model can at times ignore the benefits of an administrative process for establishing boundaries. As the analysis in Part III of this article indicates, however, the extensive grant of discretion provides abstract benefits that pale in comparison to the clear and concrete costs imposed.[112]

Responding to these problems by eliminating commissions entirely should be a non-starter. Cutting the commission process would involve returning power to political actors, thereby abbreviating the “electoral boundary revolution”[113] that took away Parliament’s power to set the map. Such a move would be misguided, as it would potentially reintroduce a risk of gerrymandering. Providing expansive discretion or eliminating it entirely, however, are not the only options. Commission discretion should instead be restructured to further democratic values such as representation by population and minority representation. More carefully calibrated constraints on commission behaviour than those currently on offer are necessary, even while preserving scope for decision making that takes into account local conditions.

A.  Rebuilding the EBRA: A Hierarchy of Criteria

Many of the problems emerging from the exercise of commission discretion stem from the structure of the EBRA. As representation by population is to be the primary criteria, the EBRA can reasonably be taken to mean that commissions should prioritize ridings of relatively equal size and that other considerations should fall if they undermine this primary goal. At the other extreme, commissions could interpret the EBRA as permitting any deviation within the twenty-five per cent range. There are grounds in the EBRA to anchor these and other interpretations. The EBRA represents what can be termed a “balancing of criteria” framework. To return to Dworkin’s categories of discretion, the balancing of criteria framework facilitates the transformation of what was intended to be weak discretion into something more, which fits less easily with an administrative body such as a boundary commission.

One alternative that has the potential to encourage a more consistent realization of the right to vote can be drawn from the “hierarchy of criteria” model applied in the United States. The bodies tasked with redistricting in many American states are obliged to follow a more defined hierarchy of criteria than found in the EBRA. On most aspects of redistricting, the United States does not form a model to emulate, but instead a cautionary tale to be avoided. The prevalence of gerrymandering and attempts at minority vote dilution mar the legitimacy of aspects of the American model. A hierarchy of criteria approach, however, stands as a useful alternative to the ERBA’s balancing model, because, if applied in a manner that fits the Canadian context, it can bring greater consistency to the exercise of commission discretion and therefore further the equality of voters.

State institutions establish both state and federal electoral districts in the United States.[114] To oversee state redistricting institutions, courts in the various states have developed hierarchies of criteria. The development of hierarchies in the jurisprudence was a functional response to the need for clear criteria for redistricting institutions, which faced conflicting sets of legal rules from state and federal sources. The hierarchies adopted by state courts incorporate the criteria stemming from the federal constitution, federal statutes such as the VRA, state constitutional rules, and state legislation in order to guide map-making bodies. These hierarchies constrain but they also enable. They limit the range of decision making by clearly establishing which factors must be taken into account in redistricting and which should yield when two principles conflict with one another. Yet they enable boundary drawers by establishing broad criteria that facilitate the exercise of weak-form discretion. The hierarchies still provide leeway for local circumstances to be taken into account. This model has the potential to restructure the discretion provided to Canadian federal commissions so that it is in conformity with the weak form that is appropriate.

The specific hierarchy adopted in each state varies, especially as state constitutions differ significantly.[115] While federal rules apply country-wide and are paramount, each state is free to develop its own set of criteria through amendment to its state constitution or legislation, though these will be lower down in the pecking order than federal ones.[116] It is therefore appropriate to speak of multiple hierarchies of criteria across the states, rather than a single one. As a general rule, however, redistricting institutions in the states must first apply the federal constitution and relevant federal quasi-constitutional legislation (namely the VRA), before looking to other criteria established at the state level. These federal criteria prioritize representation by population (or one person, one vote as it is called in the United States) and a specific approach to minority representation that results in the creation of majority-minority districts for African-American and Latino voters.[117]

Redistricting in the United States is generally carried out by the state legislature, in contrast to the Canadian commission model. There has been no equivalent to the “electoral boundary revolution” to transform the institution responsible for redistricting from the likely partisan-minded legislature to some impartial body. Commissions are a growing phenomenon in the United States, but are still used in a minority of states.[118]

The hierarchy of criteria model should not be dismissed because of these institutional differences. States that use redistricting commissions also apply the hierarchy of criteria. The exact content of the hierarchy and placement of various criteria that commissions must consider will likely be different for Canadian federal districts than for American states, which differ even amongst one another. Yet adopting this element of the American model would bring commissions closer to truly exercising weak-form discretion. I do not propose a specific hierarchy here, but leave that to another day, apart from providing general examples to illuminate how it would operate. In Canada, a hierarchy of criteria approach should be implemented through legislation, given the very small number of legal challenges on redistricting in Canada. The courts would have a role in enforcement.

In Colorado, which uses a boundary commission, a hierarchy has been clearly established in the case law.[119] These criteria ensure “equal protection for the right to participate in the … political process and the right to vote.”[120] From the top to the bottom of the hierarchy, the relevant criteria are equal protection under the Fourteenth Amendment, the right of racial minorities not to have their votes abridged on the basis of race pursuant to the Fifteenth Amendment,[121] prohibitions on racial vote dilution in section 2 of the VRA,[122] and state constitutional guarantees of population equality,[123] the integrity of cities and counties,[124] the traditional districting criteria of contiguity and compactness,[125] and protection for communities of interest.[126]

In another state that uses a commission, Arizona,[127] the criteria[128] to be taken into account for redistricting are similar to those in Colorado, with two major differences. First, Arizona gives community of interest concerns relatively higher priority. Second, the Arizona Independent Redistricting Commission is required to foster competitive districts as long as there is “no significant detriment to the other goals.”[129] The approaches in Colorado and Arizona clearly delineate which principles have priority over one another and which must yield if several conflict.

A hierarchy of criteria has several clear advantages over the balancing model deployed in the EBRA. First, imposing a hierarchy clarifies what criteria can and must be taken into account. There is enduring confusion among Canadian redistricting commissions about which criteria can be applied. The 2014 Ontario commission’s strong application of a principle that federal districts should not cross municipal boundaries, for example, raised the issue of whether administrative subunits are relevant to consider and, if so, what weight they should be given. Municipal or provincial electoral boundaries, or those of other administrative zones such as school districts, are not listed as factors to consider in the EBRA. They can be potentially considered as components of the community of interest criterion, but many commissions have ignored the electoral districts of other orders of government or administrative boundaries entirely. Including consideration of political or administrative subunits in the hierarchy, or not, would eliminate a great deal of confusion and bring greater consistency. Colorado and Arizona clearly included municipal and county boundaries, though these are lower down in the priority list.

Another criterion where clarity would be welcome is anticipated population growth. The only significant change since 1964 to the EBRA’s list of criteria to balance is that the commissions were initially directed to consider anticipated population growth. Population growth is uneven in Canada. The expectation is that urban and suburban ridings will expand much faster than rural ones in between redistributions. By building urban ridings with smaller populations in anticipation of future growth, commissions could decrease variances in population size over the course of the ten years until the next redistribution. At the behest of rural MPs who objected to less populous urban ridings than there would be otherwise, the EBRA was amended in 1975 to cut out this factor.[130]

Whether anticipated population growth could still be considered as part of the goal of achieving populations “as close as reasonably possible” to voter parity, or whether the removal of that factor from the EBRA precludes that possibility, is an open question of statutory interpretation. The legislative intent would appear to be that it should not be considered. Yet there is ambiguity because of the requirement that ridings have equal populations to the extent “reasonably possible.”[131] It is unclear on the face of the legislation whether commissions should aim for equal populations at the time they draw the maps or at some later date. Australian commissions, for example, are required to aim for population equality at the mid-point of their seven year redistricting cycle.[132]

Incorporating anticipated population growth is controversial because it would result in urban and suburban ridings with lower populations at the time of the redistribution than slower growing, rural, and remote districts. The variance permitted in the EBRA would be used in favour of those populations that have traditionally been disadvantaged by it. It is arguably fairer to consider anticipated population growth than not to, because it reduces the massive variances between riding populations that occur at the end of the ten-year cycle. Urban and suburban ridings have lower voting power because of their large populations at the time the redistribution happens, but this only worsens if population continues to grow unevenly.[133] While anticipating population growth at the riding level is not an exact science, it can be a useful tool in the service of voter equality. That the “905” suburbs around Toronto would grow, or the areas surrounding Vancouver and Calgary, for example, has been no secret over recent years. Whether anticipated population growth is included in the hierarchy of criteria, clarity would be welcome on whether commissions must consider it and how much importance they should place on it relevant to other factors.

Second, a hierarchy of criteria would foster greater coherence in the enjoyment of the right to vote across the country. This model would assist in fostering consistency in the approaches adopted by boundary commissions. The goal is not uniformity or the elimination of commission discretion. Yet if the right to vote is to be realized, then similarly situated voters should be treated similarly, regardless of whether it is the Nova Scotia commission that sets boundaries or the Manitoba commission. The criteria to be applied and the relationship among them would be consistent across commissions for federal districts. The hierarchy of criteria model does restrict the range of approaches available to commissions. No longer would one commission be able to take minority representation into account, assuming this is part of the hierarchy, while another ignored it entirely. This narrowing is appropriate, however, in order to ensure that commissions act in a way consistent with weak-form discretion and the equality of voters.

There would still inevitably be some variation, especially given the lack of case law clarifying specific redistricting terms. The list of criteria adopted in American states is not so specific as to forestall reasonable disagreement about where electoral boundaries should be placed. A hierarchy shifts the debate, however, to the application of accepted principles rather than disputes about which principles are relevant. The debate becomes one about how weak-form discretion should be applied, rather than on the selection of factors to apply typical of strong-form discretion, which is closer to what happens in practice.

Incorporating a hierarchy cannot be expected to prevent all problems regarding interpretation of redistricting criteria either. There will still be disputes about the meaning of specific criteria, particularly those that are closely related. A recent case in Colorado turned on whether a commission’s actions in a given district engaged the protections for minority Latino voters in the Voting Rights Act, which is higher up the hierarchy, or community of interest concerns, which exist lower down the priority list.[134] There is obviously some overlap between the principle of fair minority representation and recognition of communities of interest. Such disputes would be likely to occur in Canada as well. A hierarchy, however, clarifies that commissions do not have the liberty to declare that it is impossible to balance all relevant criteria and then simply ignore those they disfavour.

A hierarchy would leave less room for commissions to act on their own to develop innovative solutions to problems of democratic representation. The target set by the New Brunswick commission of a ten per cent variance, for example, was an inventive way to ensure greater adherence to representation by population. The 2004 New Brunswick commission achieved more to reform the process of electoral boundary drawing than Parliament had done in the forty years from the advent of commissions in 1964. While a hierarchy would clarify what criteria to apply and in what order, it would restrict commissions’ autonomy to develop new ways of realizing democratic values. This tradeoff is legitimate, however, given the harm caused by the divergent approaches chosen by the commissions.

B.  Changing the Commission Structure

The problem of unequal treatment of some aggregations of voters is compounded by the decentralized, federal structure of the process.[135] The use of one commission for each province, in combination with extensive discretion in balancing the competing principles in the EBRA, means that the differential treatment of voters across the country is not just a possibility, but highly likely. Even the imposition of a hierarchy of criteria would still be subject to multiple interpretations by each of the different provincial commissions. It is worthwhile considering whether the structure of the commissions could be amended to ensure the right to vote is enjoyed more evenly. I propose here to change the membership of commissions in order to advance the goal of the equal treatment of voters.

A single commission responsible for redistricting across the country was proposed by Prime Minister Diefenbaker prior to the adoption of the decentralized commission model in 1964.[136] The logic was that it would prevent having ten potentially competing interpretations of the EBRA. Given the adoption of the decentralized model, the role of the office of the Representation Commissioner was originally to sit on each commission to provide some consistency. The elimination of that position in the move from four to three member commissions in the 1970s further eroded the chances of a stable interpretation of the EBRA across the country. One way to bring the right to vote back together again would be to move from a decentralized federal approach to a single national commission that would establish maps for all provinces, as Diefenbaker originally proposed. Moving to a national commission would reflect that some of the initial concerns with having ten provincial commissions—inconsistency and unequal treatment of voters—were valid.

The benefits of maintaining ten separate commissions, however, outweigh the detriments. The commissioners in each province bring extensive local knowledge to the decisions regarding electoral maps. These local roots bring legitimacy to the work of commissions in the face of public scrutiny and consultation. A national commission would lose local knowledge and perhaps lack legitimacy in particular regions of the country or among demographic segments given the origins and backgrounds of the commissioners. Having multiple commissions may also reduce the possibility that partisans can capture the process.[137]

An alternative measure that would bring added consistency to the decision making of commissions, without a significant loss of local knowledge and legitimacy, would be to reintroduce the post of Representation Commissioner through amendments to the EBRA. The Representation Commissioner was in place for the first commissioned redistricting, but was eliminated after that point.[138] The Commissioner sat on each of the ten provincial commissions, adding a constant presence to each commission that would seem likely to increase the consistency across provincial maps. As long as the Commissioner was a non-partisan and independent Elections Canada official, drastically conflicting approaches by commissions would be less likely. The post could perhaps be similar in status to the Commissioner of Elections Canada who investigates and prosecutes electoral fraud.[139] Prior to the amendments in the Fair Elections Act,[140] the Commissioner was appointed by the Chief Electoral Officer. Putting the decision in the hands of the Chief Electoral Officer would ensure that independence and impartiality are brought to the process. The Representation Commissioner could either serve as a fourth member of each commission, or replace one of the two appointees of the Speaker. Such details could be worked out in consultation with Elections Canada. The key point is that a common member across all commissions would decrease the likelihood of arbitrary treatment of voters across provinces.

Conclusion

The move to redistricting by commissions in 1964 has justifiably earned the moniker of an “electoral boundary revolution.” Eliminating gerrymandering was an impressive feat given how much that practice had been a staple of Canadian politics since Confederation. The model adopted in 1964, however, has come under significant pressure given the conflicting approaches to democratic representation chosen by electoral boundary commissions operating with expansive discretion. The statutory and constitutional frameworks, along with the decentralized commission structure, have facilitated this proliferation of competing approaches. The result has been a fracturing of the right to vote. Commissions differ significantly in how they interpret the core principles of redistricting: representation by population, minority representation, and recognition of communities of interest. The right to vote is as a consequence enjoyed unevenly by Canadians. Reforming commission discretion to ensure equal treatment of voters across the country is a necessary step to fulfill the promise of effective representation and the principle of political equality. The Fair Representation Act dramatically altered the formula for assigning seats in the House of Commons to the provinces. How the boundaries of those ridings are drawn within each province, however, remains untouched. It is time for a similar reconsideration of electoral boundary design.

                       

[1]     SC 2011, c 26, s 2.

[2]     John C Courtney, Commissioned Ridings: Designing Canada’s Electoral Districts (Montreal: McGill-Queen’s University Press, 2001) at 4 [Courtney, Commissioned Ridings].

[3]     A note on terminology: “Redistribution” and “electoral boundary readjustment” are the traditional Canadian terms. Redistribution means the allocation of the number of electoral districts to which each province and territory is entitled. Redistribution is called “apportionment” in the United States. Electoral boundary readjustment refers to the establishment of specific geographic boundaries for electoral districts within a province. The American term is “redistricting”. I prefer the terms redistribution and redistricting for ease of use and prevalence in the literature.

[4]     There has been a revival of academic interest in electoral boundaries in recent years (see e.g. Michael Pal & Sujit Choudhry, “Is Every Ballot Equal? Visible-Minority Vote Dilution in Canada” (2007) 13:1 IRPP Choices 1 [Pal & Choudhry, “Is Every Ballot Equal?”]; Brian Studniberg, “Politics Masquerading as Principles: Representation by Population in Canada” (2009) 34:2 Queen’s LJ 611; Ron Levy, “Regulating Impartiality: Electoral-Boundary Politics in the Administrative Arena” (2008) 53:1 McGill LJ 1; Andrew Sancton, The Principle of Representation by Population in Canadian Federal Politics (Toronto: Mowat Centre for Policy Innovation, 2010), online: <mowatcentre.ca/wp-content/uploads/publications/2_the_principle_of_representation.pdf>; Matthew Mendelsohn & Sujit Choudhry, Voter Equality and Other Canadian Values: Finding the Balance (Toronto: Mowat Centre for Policy Innovation, 2011), online: <mowatcentre.ca/wp-content/uploads/publications/35_voter_equality_and_other.pdf>; Mark Carter, “Ambiguous Constitutional Standards and the Right to Vote” (2011) 5 JPPL 309 [Carter, “Ambiguous Constitutional Standards”]; Michael Pal & Sujit Choudhry, “Still Not Equal? Visible Minority Vote Dilution in Canada” (2014) 8:1 Can Political Science Rev 85 [Pal & Choudhry, “Still Not Equal?”]).

[5]     I consider the distribution of seats among the provinces in Michael Pal, “Fair Representation in the House of Commons?” (2015) Special Issue JPPL 35. See also Lori Turnbull, “The Implications of the Fair Representation Act for the Parliament of Canada” (2013) 7 JPPL 25; Hugh Mellon & Don Kerr, “The Fair Representation Act of 2012 Fully Explained” (2012) 6 JPPL 545.

[6]     See Courtney, Commissioned Ridings, supra note 2 at 11, 20–23.

[7]     RK Carty, “The Electoral Boundary Revolution in Canada” (1985) 15:3 Am Rev Can Stud 273.

[8]     See Courtney, Commissioned Ridings, supra note 2 at 74.

[9]     The need for further reforms was forcefully put by John C. Courtney in the 1980s (“Parliament and Representation: The Unfinished Agenda of Electoral Redistributions” (1988) 21:4 Can J Political Science 675 [Courtney, “Unfinished Agenda”]).

[10]         RSC 1985, c E-3 [EBRA].

[11]         See generally Samuel Issacharoff, “Gerrymandering and Political Cartels” (2002) 116:2 Harv L Rev 593.

[12]         See generally Norman Ward, “The Basis of Representation in the House of Commons” (1949) 15:4 Can J Economics & Political Science 477 at 489, 491–92; Courtney, Commissioned Ridings, supra note 2 at 20–21, 23.

[13]         Courtney, Commissioned Ridings, supra note 2 at 20.

[14]         See ibid at 22–23.

[15]         See ibid at 57–62.

[16]         See ibid at 67, 96.

[17]         See Levy, supra note 4 at 54.

[18]         See Courtney, Commissioned Ridings, supra note 2 at 94–121 (detailing the composition of commissions).

[19]         Ronald Dworkin, Taking Rights Seriously (Cambridge, Mass: Harvard University Press, 1978) at 31. For more on this point, see ibid at 31–39.

[20]         Ibid at 31–33. The various iterations of the Hart-Dworkin debate have engaged with the distinction between weak and strong discretion, particularly with regard to judicial discretion (see e.g. Nicola Lacey, “The Path Not Taken: H. L. A. Hart’s Harvard Essay on Discretion” (2013) 127:2 Harv L Rev 636 at 643–44; Brian Leiter, “Beyond the Hart/Dworkin Debate: The Methodology Problem in Jurisprudence” (2003) 48 Am J Juris 17 at 20–21; Michael Giudice, “Existence and Justification Conditions of Law” (2003) 16:1 Can JL & Jur 23 at 35, n 46). I do not intend to take any position with regard to the debate between Dworkin and legal positivists here. Dworkin’s approach to categorizing types of discretion has been used in the election law context as well (see James M Fischer, “Discretion and Politics: Ruminations on the Recent Presidential Election and the Role of Discretion in the Florida Presidential Election Recount” (2001) 69:3 U Cin L Rev 807). For an application of the categories in Canadian administrative law, see Susan L Gratton, “Standing at the Divide: The Relationship Between Administrative Law and the Charter Post-Multani” (2008) 53:3 McGill LJ 477 at 482.

[21]         Dworkin, supra note 19 at 31.

[22]         Ibid at 32.

[23]         See ibid.

[24]         Ibid at 32–33.

[25]         (Cth) [CEA]. See also Graeme Orr, The Law of Politics: Elections, Parties and Money in Australia (Sydney: Federation Press, 2010) at 31, 39–44.

[26]         (UK), c 56. The Act was amended recently by the Parliamentary Voting System and Constituencies Act 2011 (UK), c 1 [PVSCA].

[27]         Supra note 10, s 15(1)(a).

[28]         Ibid, s 15(1)(b).

[29]         Ibid, s 15(2).

[30]         Ibid.

[31]         The United States famously requires exact population equality for congressional districts. For classic articulations of this principle, see Baker v Carr, 369 US 186, 82 S Ct 691 (1962); Wesberry v Sanders, 376 US 1, 84 S Ct 526 (1964); Reynolds v Sims, 377 US 533, 84 S Ct 1362 (1964). The jurisprudence does permit somewhat larger deviations for state districts than for federal ones (see e.g. Brown v Thomson, 462 US 835, 103 S Ct 2690 (1983) in which the court established a ten per cent variance in certain circumstances for state legislative districts). However, the United States Supreme Court imposed a standard closer to one person one vote in the face of partisan gerrymandering (see Cox v Larios, 542 US 947, 124 S Ct 2806 (2004); Samuel Issacharoff & Pamela S Karlan, “Where to Draw the Line?: Judicial Review of Political Gerrymanders” (2004) 153:1 U Pa L Rev 541). There is controversy about the value of this standard in the literature (see e.g. Grant M Hayden, “The False Promise of One Person, One Vote” (2003) 102:2 Mich L Rev 213). Another geographically large federation, Australia, makes do with a ten per cent permissible variance (see CEA, supra note 25, s 66(3)). Italy applies a ten per cent variance as well, while New Zealand uses five per cent and Germany fifteen per cent (see Lisa Handley, “Challenging the Norms and Standards of Election Administration: Boundary Delimitation” in Challenging the Norms and Standards of Election Administration (IFES, 2007) 59 at 63, n 8). The United Kingdom’s recently passed legislation imposes a five per cent variance (see PVSCA, supra note 26, s 11(1)(2)(1)) and, in lieu of an extraordinary circumstances clause, lists the small number of districts that are exempt from the strict limit (see ibid, s 11(1)(6)(3)). These exempt constituencies are mainly remote islands off the coast of Scotland (see ibid).

[32]         Canadian provinces have moved in the direction of voter equality, with Saskatchewan, Manitoba, New Brunswick, and Newfoundland and Labrador all implementing variances lower than those in place federally. See The Electoral Divisions Act, CCSM, c E40, s 11(3) (applying a ten per cent variance in Manitoba with exceptions for northern constituencies); The Constituency Boundaries Act, 1993, SS 1993, c C-27.1, ss 14(3)–(4) (setting a five per cent standard in Saskatchewan, though with exemptions for two northern constituencies); Electoral Boundaries Act, RSNL 1990, c E-4, ss 15(2), 15(6) (establishing a ten per cent variance in Newfoundland with exemptions for Labrador); Electoral Boundaries and Representation Act, RSNB 2014, c 106, ss 11(5)–(6) (using fifteen per cent or no more than twenty-five per cent in extraordinary circumstances in New Brunswick). See also Mendelsohn & Choudhry, supra note 4 at 11.

[33]         Canadian Charter of Rights and Freedoms, Part I of the Constitution Act, 1982, being Schedule B to the Canada Act 1982 (UK), 1982, c 11 [Charter].

[34]         See Reference Re Provincial Electoral Boundaries (Sask), [1991] 2 SCR 158, 81 DLR (4th) 16 [Carter cited to SCR].

[35]         See Figueroa v Canada (Attorney General), 2003 SCC 37, [2003] 1 SCR 912 [Figueroa]. See also Heather MacIvor, “The Charter of Rights and Party Politics: The Impact of the Supreme Court Ruling in Figueroa v. Canada (Attorney General)” (2004) 10:4 IRPP Choices 1.

[36]         Supra note 34.

[37]         There is a convincing case that the electoral map was the product of a gerrymander by an incumbent government. The Devine government benefitted from the votes of rural dwellers and therefore sought to overrepresent them to its own partisan advantage and to the disadvantage of the opposing parties which had stronger urban support. See Michael Pal, “Breakdowns in the Democratic Process and the Law of Canadian Democracy” (2011) 57:2 McGill LJ 299 at 301, 335 [Pal, “Breakdowns”]; Carter, “Ambiguous Constitutional Standards”, supra note 4 at 320–21; Yasmin Dawood, “Electoral Fairness and the Law of Democracy: A Structural Rights Approach to Judicial Review” (2012) 62:4 UTLJ 499 at 537–44; Christopher D Bredt & Markus F Kremer, “Section 3 of the Charter: Democratic Rights at the Supreme Court of Canada” (2004) 17 NJCL 19 at 22; James M Pitsula & Kenneth A Rasmussen, Privatizing a Province: The New Right in Saskatchewan (Vancouver: New Star Books, 1990) at 254–55; David Johnson, “Canadian Electoral Boundaries and the Courts: Practices, Principles and Problems” (1994) 39:1 McGill LJ 224 at 228; Robert G Richards & Thomson Irvine, “Reference Re Provincial Electoral Boundaries: An Analysis” in John C Courtney, Peter MacKinnon & David E Smith, eds, Drawing Boundaries: Legislatures, Courts, and Electoral Values (Saskatoon: Fifth House, 1992) 48 at 51, 62. For more on the allegations of gerrymandering, see also Kent Roach, “Chartering the Electoral Map into the Future” in Courtney, MacKinnon & Smith, supra note 37, 200 at 207; Kent Roach, “One Person, One Vote? Canadian Constitutional Standards for Electoral Distribution and Districting” in David Small, ed, Drawing the Map: Equality and Efficacy of the Vote in Canadian Electoral Boundary Reform (Toronto: Dundurn Press, 1991) 3 at 10, 28, 40–41; Janet Hiebert, “Representation and the Charter: Should Rights Be Paramount?” in Courtney, MacKinnon & Smith, supra note 37, 1 at 6, 14.

[38]         Carter, supra note 34 at 183. For criticism of the reasoning in Carter, see Pal, “Breakdowns”, supra note 37 at 301–02, 335; Studniberg, supra note 4 at 638–45; Carter, “Ambiguous Constitional Standards”, supra note 4 at 320–22; Bredt & Kremer, supra note 37 at 23, 53–59.

[39]         Carter, supra note 34 at 184.

[40]         See Bredt & Kremer, supra note 37 at 22.

[41]         John Courtney found that immediately after Carter, boundary commissions in the 1996 redistribution made frequent reference to the decision (Commissioned Ridings, supra note 2 at 172–203). The actual impact of Carter in standardizing commission behaviour, I will argue, was negligible.

[42]       Australia, for example, adopted a similarly permissive regime with regard to departing from population equality (see McGinty v Western Australia, [1996] HCA 48, 186 CLR 140; Attorney-General (Cth), Ex rel McKinlay v Commonwealth [1975] HCA 53, 135 CLR 1).

[43]         The claim in Carter, supra note 34, that accepting large deviations from representation by population was a principled decision in the history of redistricting ignores the extensive evidence in Canada of gerrymandering (see Courtney, Commissioned Ridings, supra note 2 at 20–21; Courtney, “Unfinished Agenda”, supra note 9 at 686; John C Courtney, “‘Theories Masquerading as Principles’: Canadian Electoral Boundary Commissions and the Australian Model” in John C Courtney, ed, The Canadian House of Commons: Essays in Honour of Norman Ward (Calgary: University of Calgary Press, 1985) 135; Carty, supra note 7 at 283; Ward, supra note 12 at 489; Ronald E Fritz, “The Saskatchewan Electoral Boundaries Case and Its Implications” in Courtney, MacKinnon & Smith, supra note 37, 70 at 74–77). Fritz writes: “[p]erhaps [Justice McLachlin] considered this description of Canadian constitutional history as being without controversy. … However, this view … represents only a superficial account, giving a patina of respectability to the true political motivations that drove the historical developments” (ibid at 74).

[44]         See Frank v Canada (AG), 2015 ONCA 536, 388 DLR (4th) 1 (limiting the right of non-resident citizens living abroad for five years or more to cast ballots, but on the basis of section 1). On appeal, the federal government abandoned the argument that it had made in the lower court that residency was an internal limit on section 3 rights. At trial, the court had decisively rejected the internal limit claim (see Frank v Canada (AG), 2014 ONSC 907 at para 79, 372 DLR (4th) 681). The Superior Court points out that the drafters of the Charter specifically included only citizenship in the text as a qualifier on the right to vote. It can therefore be reasonably inferred that other potential internal limits, such as residence or age, should not be read into the provision (ibid at para 84). The Supreme Court has granted leave to appeal in Frank (36645 (14 April 2016)). See also Haig v Canada (Chief Electoral Officer), [1993] 2 SCR 995, 105 DLR (4th) 577 (upholding residency requirements for a Quebec referendum, but finding that section 3 is inapplicable to referenda and that there is nothing in section 2(b) that requires open participation by all in a referendum in a specific province).

[45]         Sauvé v Canada (Chief Electoral Officer), 2002 SCC 68 at para 14, [2002] 3 SCR 519.

[46]         See ibid.

[47]         See Canadian Disability Rights Council v Canada, [1988] 3 FC 622, 21 FTR 268 (FCTD).

[48]         See Muldoon v Canada, [1988] 3 FC 628, 21 FTR 154 (FCTD).

[49]         See Libman v Quebec (AG), [1997] 3 SCR 569, 151 DLR (4th) 385.

[50]         See Harper v Canada (AG), 2004 SCC 33, [2004] 1 SCR 827.

[51]         Ibid at para 62. See also Colin Feasby, “Libman v. Quebec (A.G.) and the Administration of the Process of Democracy under the Charter: The Emerging Egalitarian Model” (1999) 44:1 McGill LJ 5; Colin Feasby, “Freedom of Expression and the Law of the Democratic Process” (2005) 29 SCLR (2d) 237 at 263–64, 277–82; MacIvor, supra note 35 at 5–6. On the egalitarian model more generally, see Frederick Schauer, “Judicial Review of the Devices of Democracy” (1994) 94:4 Colum L Rev 1326 at 1340–41. For critical takes on the egalitarian model, see Christopher D Bredt & Laura Pottie, “Liberty, Equality and Deference: A Comment on Colin Feasby’s ‘Freedom of Expression and the Law of the Democratic Process’” (2005) 29 SCLR (2d) 291 at 292–99; Christopher Manfredi & Mark Rush, Judging Democracy (Peterborough: Broadview Press, 2008) at 97–117; Jay Makarenko, “Fair Opportunity to Participate: The Charter and the Regulation of Electoral Speech” (2009) 3:2 Can Political Science Rev 38.

[52]         Supra note 49.

[53]         2007 SCC 12, [2007] 1 SCR 527. For a discussion of this case, see Michael Pal, “Democratic Rights and Social Science Evidence” (2014) 32:2 NJCL 151 [Pal, “Social Science Evidence”].

[54]         SC 2000, c 9.

[55]         Despite its constitutionality, the Chief Electoral Officer recommended the provision be scrapped (see Elections Canada, Report of the Chief Electoral Officer of Canada on the 41st General Election of May 2, 2011 (Ottawa: Elections Canada, 2011) at 49) and Parliament eventually repealed it (see Bill C-23, An Act to amend the Canada Elections Act and other Acts and to make consequential amendments to certain Acts, 2nd Sess, 41st Parl, 2014 (assented to 14 June 2014) [Fair Elections Act]). See also Pal, “Social Science Evidence”, supra note 53 at 161, n 64.

[56]         Political parties had to field fifty or more candidates in order to be registered. Registration brought with it the ability to issue tax receipts outside of the election period, the right of candidates to transfer unspent election funds to the party, and the right to list a candidate’s party affiliation on the ballot (see Figueroa, supra note 35 at para 4).

[57]         The result is the endorsement of a party equality model (see MacIvor, supra note 35 at 5 for a description of the term). The protection for small political parties was extended to independent candidates in R v Nunziata, 2005 ONCJ 292, 78 OR (3d) 285. Longley v Canada (AG), 2007 ONCA 852, 88 OR (3d) 408 reduced the reach of the party equality principle by upholding as constitutional thresholds for the receipt of the per vote quarterly allowance that harmed small parties. The quarterly allowance has now been phased out (see Keeping Canada’s Economy and Jobs Growing Act, SC 2011, c 24, s 181).

[58]         Carter, supra note 34 at 184.

[59]         See Jennifer Smith & Ronald G Landes, “Entitlement Versus Variance Models in the Determination of Canadian Electoral Boundaries” (1998) 17 Intl J Can Stud 19; Courtney, Commissioned Ridings, supra note 2 at 221, 225–31.

[60]         Mixed member proportional electoral systems have some representatives elected directly with others chosen from a party list. This practice raises questions about whether there are two classes of representatives within the same legislative body, in terms of their roles and even legitimacy. For detailed discussion on this type of electoral system, see Matthew Soberg Shugart & Martin P Wattenberg, eds, Mixed-Member Electoral Systems: The Best of Both Worlds? (New York: Oxford University Press, 2001); Leigh J Ward, “‘Second-Class MPs’? New Zealand’s Adaptation to Mixed-Member Parliamentary Representation” (1998) 49:2 Political Science 125.

[61]         For an explanation of these constitutional rules, see Pal & Choudhry, “Is Every Ballot Equal?”, supra note 4 at 12.

[62]         See Federal Electoral Boundaries Commission for Alberta, Report of the Federal Electoral Boundaries Commission for the Province of Alberta: 2012 (Ottawa: Elections Canada, 2012) at 8, 26, online: <www.redecoupage-federal-redistribution.ca/
ab/now/reports/ab_report_e.pdf> [Alberta Commission, Report].

[63]         Whatever approach they adopt, commissions generally stick more closely to representation by population in the initial maps that they propose. These initial maps are the basis for province-wide public consultations. Public submissions inevitably push for less adherence to representation by population and promote instead the preservation of community boundaries. Commissions tend, therefore, to bow to this pressure and put in place final maps that deviate to a greater extent from representation by population than in the initial maps (see Courtney, Commissioned Ridings, supra note 2 at 135–37).

[64]         These were Yellowhead and Red Deer-Wolf Creek (see Alberta Commission, Report, supra note 62 at 55–77).

[65]         Courtney, Commissioned Ridings, supra note 2 at 178.

[66]         See Federal Electoral Boundaries Commission for New Brunswick, Report of the Federal Electoral Boundaries Commission for the Province of New Brunswick: 2012 (Ottawa: Elections Canada, 2012) at 6, online: <www.redecoupage-federal-redistribution.ca/
nb/now/reports/report_e.pdf>.

[67]         See Federal Electoral Boundaries Commission for New Brunswick, Report of the Federal Electoral Boundaries Commission for the Province of New Brunswick (Ottawa: Elections Canada, 2003) [New Brunswick Commission, Report].

[68]       Lower variance targets appear to have been reflected in 2014 in British Columbia and Alberta’s initial maps, though the commissions do not frame it that directly (see Federal Electoral Boundaries Commission for British Columbia, Electoral Districts: Proposal for the Province of British Columbia (Ottawa: Elections Canada, 2014) at 19–20, online: <www.redecoupage-federal-redistribution.ca/bc/now/proposals/bc_proposal_e.pdf>; Federal Electoral Boundaries Commission for Alberta, Electoral Districts: Proposal for the Province of Alberta (Ottawa: Elections Canada, 2014) at 26–27, online: <www.
redecoupage-federal-redistribution.ca/ab/now/proposals/ab_proposal_e.pdf>).

[69]         2004 FC 679, [2005] 1 FCR 93 [Raîche].

[70]       How to incorporate minority representation for Acadians has proved a challenge (see Charlottetown v Prince Edward Island (1998), 169 Nfld & PEIR 188 at paras 89, 95, 168 DLR (4th) 79 (PEI SC); Smith & Landes, supra note 59 at 24, 27). For more recent coverage, see “New Brunswick Acadian Group Upset with New Provincial Electoral Map”, Times Colonist (25 April 2013), online: <www.timescolonist.com/life/new-brunswick-acadian-group-upset-with-new-provincial-electoral-map-1.138932>. For a related controversy in Nova Scotia, see David Jackson, “Acadians to File Legal Challenge of Proposed Electoral Boundaries”, The Chronicle Herald (4 December 2012), online: <www.thechronicleherald.ca/novascotia/217712-acadians-to-file-legal-challenge-of-proposed-electoral-boundaries>.

[71]   The map moved the parish of Allardville and parts of the parishes of Saumarez and Bathurst from the Acadie-Bathurst riding to the adjacent Miramichi riding (see New Brunswick Commission, Report, supra note 67).

[72]         Critics disagreed with the move to equalize the populations between the two districts. A petition and presentations made on behalf of francophone voters to the boundary commission during the public consultation objected to the change. The Official Languages Commissioner and the House of Commons Standing Committee on Procedure and House Affairs, the latter acting at the behest of Acadie-Bathurst MP Yvon Godin, both recommended that the commission preserve the entire francophone presence within Acadie-Bathurst (see Canada, Office of the Commissioner of Official Languages, Drawing the Line: The Impact of Readjusting the Electoral Boundaries on the Official Language Minority Communities (Ottawa: Minister of Public Works and Government Services Canada, 2006) at 11). The boundary commission modified its initial proposal in the direction urged by the Official Languages Commissioner and the House Committee (see Federal Electoral Boundaries Commission for the Province of New Brunswick, Disposition of the Commission Pursuant to Subsection 23(1) of the Electoral Boundaries Readjustment Act of Objections Filed by Members of the House of Commons with Respect to the Commission’s Report Dated January 10, 2003 (Ottawa: Elections Canada, 2003) at 5–6 [New Brunswick Commission, Disposition]). It moved the parish of Saumarez and part of the parish of Allardville back to Acadie-Bathurst, but kept Bathurst and the remaining portion of Allardville in the Miramichi riding (ibid). Some Acadians previously within francophone-majority Acadie-Bathurst would still have found themselves in a francophone-minority district under the new plan.

[73]         Supra note 10, s 15(1)(b)(i). The claimants also asserted that the map violated the Charter for harming Acadian voting rights and breached the Official Languages Act, RSC 1985, c 31 (4th Supp) [OLA]. The OLA obliges federal government institutions, including the boundary commissions, to act to enhance the vitality of French and English minority language communities (ibid, s 41). Both the Charter and OLA claims were dismissed.

[74]         I thank former Supreme Court Justice Michel Bastarache for this point.

[75]         Thamotharem v Canada (Minister of Citizenship and Immigration), 2007 FCA 198 at para 78, [2008] 1 FCR 385, leave to appeal to SCC refused, 32185 (13 December 2007). See also Ha v Canada (Minister of Citizenship and Immigration), 2004 FCA 49 at paras 70–78, [2004] 3 FCR 195; Ainsley Financial Corp v Ontario (Securities Commission) (1994), 21 OR (3d) 104 at 110, 121 DLR (4th) 79 (CA); Maple Lodge Farms v Government of Canada, [1982] 2 SCR 2, 137 DLR (3d) 558.

[76]         See New Brunswick Commission, Disposition, supra note 72 at 5–6.

[77]         See Federal Electoral Boundaries Commission for Ontario, Report of the Federal Electoral Boundary Commission for the Province of Ontario (Ottawa: Elections Canada, 2012), online: <www.redecoupage-federal-redistribution.ca/on/now/reports/on_report_e.
pdf> [Ontario Commission, Report]. For the Ontario Commission’s overall approach, see ibid at 8–9; for specific deviations from representation by population, see ibid at 152–56. In critiquing the Ontario map, I draw here on Michael Pal & Melissa Molson, Moving Toward Voter Equality: Mowat Centre Report on the Proposed Federal Electoral Boundaries for Ontario (Toronto: Mowat Centre for Policy Innovcation, 2012) at 4–11, online: <mowatcentre.ca/moving-toward-voter-equality/>. The 2014 Ontario map is only one example, however, from many across the 2004 and 2014 redistribution cycles. For example, commissions in 2014 in Quebec, Newfoundland and Labrador, New Brunswick, and Nova Scotia displayed relatively little enthusiasm for representation by population.

[78]         The regions were: (1) Northern Ontario; (2) Southwestern Ontario; (3) Central South Ontario; (4) Halton, Hamilton and Niagara; (5) Georgian Bay, Barrie and Simcoe; (6) Brampton and Mississauga; (7) Newmarket, York and Vaughan; (8) City of Toronto; (9) Durham, Port Hope and Cobourg; (10) Haliburton, Peterborough and Quinte West; and (11) Eastern Ontario (see Ontario Commission, Report, supra note 77 at 9–35).

[79]         For the full list of sub-regions and a discussion, see Pal & Molson, supra note 77 at 5.

[80]         See Ontario Commission, Report, supra note 77 at 154.

[81]         See ibid at 155.

[82]         See ibid at 153.

[83]         See Pal & Choudhry, “Is Every Ballot Equal?”, supra note 4 at 6–7; Pal & Choudhry, “Still Not Equal?”, supra note 4.

[84]         See Pal & Choudhry, “Is Every Ballot Equal?”, supra note 4 at 7–8; Pal & Choudhry, “Still Not Equal?”, supra note 4.

[85]         See Ontario Commission, Report, supra note 77 at 153.

[86]         Federal Electoral Boundaries Commission for Ontario, Proposal for the Province of Ontario (Ottawa: Elections Canada, 2012) at 9, online: <www.redecoupage-federal-redistribution.ca/on/now/proposals/on_proposal_e.pdf>.

[87]         For a broader discussion of minority representation and electoral systems, see e.g. Lani Guinier, The Tyranny of the Majority: Fundamental Fairness in a Representative Democracy (New York: Free Press, 1994); Melissa S Williams, Voice, Trust, and Memory: Marginalized Groups and the Failings of Liberal Representation (Princeton: Princeton University Press, 1998); Will Kymlicka, Multicultural Citizenship: A Liberal Theory of Minority Rights (Oxford: Oxford University Press, 1995) at 132–38.

[88]         The problem of inconsistent application of community of interest concerns was pointed out as a potential problem twenty-five years ago (see Alan Stewart, “Community of Interest in Redistricting” in Small, supra note 37, 117). See also Johnson, supra note 37 at 237–39; Studniberg, supra note 4 at 639–42; John C Courtney, “Community of Interest in Electoral Boundary Readjustments” (2002) 4:2 Electoral Insight 9.

[89]         Carter, supra note 34 at 184.

[90]         Canada, Royal Commission on Electoral Reform and Party Financing, Reforming Electoral Democracy: Final Report, vol 1 (Ottawa: Supply and Services Canada, 1991).

[91]         Ibid at 158.

[92]         See Courtney, Commissioned Ridings, supra note 2 at 224–31.

[93]         See ibid at 215; Richard W Jenkins, “Untangling the Politics of Electoral Boundaries in Canada, 1993–1997” (1998) 28:4 Am Rev Can Stud 517 at 525.

[94]         See Courtney, Commissioned Ridings, supra note 2 at 215; Jenkins, supra note 93 at 525–26; Bill C-69, An act to provide for the establishment of electoral boundaries commissions and the readjustment of electoral boundaries, 1st Sess, 35th Parl, 1995, s 19(5).

[95]         Federal Electoral Boundaries Commission for Alberta, Report of the Federal Electoral Boundaries Commission for Alberta (Ottawa: Elections Canada, 2003) at 13.

[96]         Ibid at 13.

[97]         Levy, supra note 4 at 12 argues that commissions are neutral, not just with regard to partisanship, but also other interests. Carter describes Levy’s claim this way: “According to Levy, the Canadian scheme reflects a ‘more robust’ kind of impartiality which is characterized by the commissioners’ neutrality in relation to political, social, or other ideological commitments” (“Ambiguous Constitutional Standards”, supra note 4 at 313). In my view, the diversity of approaches adopted by the commissions would suggest that formal impartiality or neutrality in fact masks a plethora of commitments with regard to minority rights.

[98]         See Trevor Knight, “Electoral Justice for Aboriginal People in Canada” (2001) 46:4 McGill LJ 1063 at 1069.

[99]         To the best of my knowledge this issue was first raised by the political scientist Dennis Pilon in an interview with CBC news (see “Sask. Riding Boundaries Unique: Political Scientists”, CBC News (29 April 2011), online: <www.cbc.ca/news/canada/
saskatchewan/sask-riding-boundaries-unique-political-scientists-1.1027878>). See also Simon Enoch, “The Strange Case of Saskatchewan’s Electoral Boundaries” (9 August 2011), Behind the Numbers (blog), online: <behindthenumbers.ca/2011/08/09/the-strange-case-of-saskatchewans-electoral-boundaries/>.

[100] The New Democratic Party (NDP) has traditionally seen Saskatchewan as a stronghold, particularly its cities, and has recorded significant levels of support in recent elections, including twenty-three per cent in 2004, twenty-four per cent in 2006, twenty-six per cent in 2008, and thirty-two per cent in 2011. Despite this degree of backing, the NDP were shut out in the province in the four most recent elections. Similar levels of support saw the NDP win two seats in the 2000 election. By diluting the urban vote, which tends to vote for the NDP, with the rural vote, which tends to vote for the Conservatives, the boundary commission affected Saskatchewan’s representation in the House. See Enoch, supra note 99.

[101] Federal Electoral Boundaries Commission for Ontario, Report of the Federal Electoral Boundaries Commission for Ontario (Ottawa: Elections Canada, 2003) at 19.

[102] Benjamin Forest has done the most complete study on minority influence districts in Canada. He finds that a concentrated Asian presence in large cities facilitates the creation of districts with strong minority voting populations. See Benjamin Forest, “Electoral Redistricting and Minority Political Representation in Canada and the United States” (2012) 56:3 Can Geographer 318 at 329.

[103] Voting Rights Act of 1965, 52 USC § 10301 [VRA].

[104] See Smith & Landes, supra note 59 at 27–28.

[105] See ibid at 29–31.

[106] Levy, supra note 4.

[107] Ibid at 6, 15–23.

[108] Ibid at 17.

[109] Ibid at 19.

[110] Ibid at 20–21.

[111] See generally Pal, “Breakdowns”, supra note 37.

[112] I agree with the overall conclusion reached by Mark Carter in his critique of Levy that further constraints are preferable to the vast discretion granted to commissions. He argues that the benefits of ambiguity and extensive discretion are minimal in comparison to the harms caused by partisan gerrymandering, which occurred in Saskatchewan in the Carter case, and vote dilution (see Carter, “Ambiguous Constitutional Standards”, supra note 4 at 311). See also Mark Carter, “Reconsidering the Charter and Electoral Boundaries” (1999) 22:1 Dal LJ 53 at 58–60, 71.

[113] Carty, supra note 7.

[114] See US Const art I, § 4 (“[t]he Times, Places and Manner of holding Elections for Senators and Represenatives, shall be prescribed in each State by the Legislature thereof”).

[115] On state constitutions, see Jonathan L Marshfield, “Models of Subnational Constitutionalism” (2011) 115:4 Penn St L Rev 1151.

[116] See Re Reapportionment of the Colorado General Assembly, 332 P (3d) 108, 2011 WL 5830123 (Colo Sup Ct 2011) [Colorado 2011].

[117] The United States Supreme Court struck down section 4 of the VRA as unconstitutional in Shelby County, Ala v Holder, 133 S Ct 2612 (2013), 186 L Ed 2(d) 651. Section 4 set the formula for determining which states and counties had obligations to “pre-clear” changes to electoral laws with federal officials. Striking down section 4 in effect ended the pre-clearance obligation set out in section 5 of the VRA, but Shelby County did not address the creation of majority-minority districts. On the VRA’s impact on minority representation, see Heather K Gerken, “A Third Way for the Voting Rights Act: Section 5 and the Opt-In Approach” (2006) 106:3 Colum L Rev 708; Samuel Issacharoff, “Is Section 5 of the Voting Rights Act a Victim of Its Own Success?” (2004) 104:6 Colum L Rev 1710; Nathaniel Persily, “The Promise and Pitfalls of the New Voting Rights Act” (2007) 117:2 Yale LJ 174.

[118] See Bruce E Cain, “Redistricting Commissions: A Better Political Buffer?” (2012) 121:7 Yale LJ 1808 at 1812–21; D Theodore Rave, “Politicians as Fiduciaries” (2013) 126:3 Harv L Rev 671 at 680, n 38, 729–35; Nicholas O Stephanopoulos, “Our Electoral Exceptionalism” (2013) 80:2 U Chicago L Rev 769 at 778–80; Nicholas D Mosich, “Judging the Three-Judge Panel: An Evaluation of California’s Proposed Redistricting Commission” (2005) 79:1 S Cal L Rev 165; Kristina Betts, “Redistricting: Who Should Draw the Lines? The Arizona Independent Redistricting Commission as a Model for Change” (2006) 48:1 Ariz L Rev 171.

[119] See Colorado 2011, supra note 116; Re Reapportionment of the Colorado General Assembly, 45 P (3d) 1237, 2002 WL 100555 (Colo Sup Ct 2002) [Colorado 2002]; Re Reapportionment of the Colorado General Assembly, 828 P (2d) 185, 1992 WL 48569 (Colo Sup Ct 1992); Re Reapportionment of the Colorado General Assembly, 647 P (2d) 191 (Colo Sup Ct 1982).

[120] Colorado 2002, supra note 119 at 1241.

[121] The hierarchy places a priority on federal law. The Equal Protection Clause of US Const amend XIV, § 1 bars states from denying “to any person within its jurisdiction the equal protection of the laws.” US Const amend XV, § 1 states: “The right of citizens of the United States to vote shall not be denied or abridged by the United States or by any state on account of race, color, or previous condition of servitude.”

[122] The VRA prohibits a state from imposing a “voting qualification or prerequisite … in a manner which results in a denial or abridgement of the right of any citizen … to vote on account of race or color” (supra note 100, USC § 10301).

[123] See Colo Const art V, § 46 which requires population equality with no more than five per cent deviation between the most and least populous district in each house.

[124] See ibid, art V, § 47(2).

[125] See ibid, art V, § 47(1).

[126] See ibid, art V, § 47(3).

[127] See Cain, supra note 118 at 1830–37; Rave, supra note 118 at 730–35; Stephanopoulos, supra note 118 at 780, 790, 803; Betts, supra note 118 at 190–94.

[128] Each criterion is to be fulfilled to the “extent practicable”, with the exception of compliance with federal law, which is not limited by that qualifier (Ariz Const art IV, pt 2, § 1(14)).

[129] Ibid.

[130] See Courtney, Commissioned Ridings, supra note 2 at 65–66.

[131] EBRA, supra note 10, s 15(1)(a).

[132] See CEA, supra note 25, s 63A(2). Australia, like Canada, has a non-partisan process and significant demographic diversity that must be addressed in drawing boundaries.

[133] See Pal & Choudhry, “Is Every Ballot Equal?”, supra note 4 at 13.

[134] See Colorado 2011, supra note 116. The state court ruled that the district engaged community of interest concerns and not compliance with the VRA. The map was therefore held to be unlawful because community of interest concerns had been placed by the commission above a rule against unnecessarily splitting counties and cities in the drawing of districts. A more persuasive argument, in my opinion, is that the commission’s map was permissible for prioritizing minority representation above concerns lower down the hierarchy.

[135] This consequence of decentralized commissions is highlighted in Courtney, “Unfinished Agenda”, supra note 9 at 676–78.

[136] See Courtney, Commissioned Ridings, supra note 2 at 58.

[137] See Levy, supra note 4 at 54–56 (discussing decentralization and impartiality).

[138] See Courtney, Commissioned Ridings, supra note 2 at 96.

[139] See Canada Elections Act, supra note 54, ss 509–21.

[140] Supra note 55.