Double Aspect

Earlier this week, the Court of Appeal for Ontario delivered its decision in Fair Voting BC v Canada (Attorney General), 2025 ONCA 581, rejecting a challenge to the constitutionality of Canada’s use of the “first-past-the-post” (or, if you want to be fancy about it, single-member plurality) voting system. The outcome is not especially suprising; the Quebec Court of Appeal ruled to the same effect in Daoust c Directeur général des élections du Québec, 2011 QCCA 1634. But Justice Huscroft’s majority reasons make a number of quite interesting points, and Justice Dawe’s brief concurrence is worth noting too. That said, I will keep this post fairly short, because I am due to comment on the case at greater length elsewhere in the coming months. (All shall be revealed in due course.) If you have seen the threads I have posted on Twitter and Bluesky, there will not be much new stuff here.

For the most part, I agree with Justice Huscroft. So let me start with one small point on which I decidedly don’t, especially since it arises out of the way in which he opens his reasons. Justice Huscroft writes:

Canada has had the same federal electoral system since 1867. Candidates representing political parties seek election in over 300 ridings across the country. The candidate who gets the most votes wins the riding, and following constitutional convention the Governor General invites the political party whose candidates win the most ridings to form a government and seek the confidence of Parliament.

Actually there are two mistakes here, one perhaps insignificant, the other important. The former has to do with the assumption that single-member plurality was always the only electoral system used in Canada. Not so. Section 40 of the the Constitution Act, 1867 provided, in part, that “[t]he County of Halifax shall be entitled to return Two Members” in the newly-created Dominion Parliament. Moreover, in a comment here Peter McCormick pointed out that two-member districts were used, not only in Halifax but elsewhere too, for the first century of Canada’s existence, and even longer in some provinces. This is mostly a piece of trivia, and is not really relevant to Justice Huscroft’s opinion, but it is a fact worth knowing because some people have, from time to time, argued that first-past-the-post is somehow embedded in the Constitution Act, 1867 or in Canada’s “constitutional architecture” and is beyond either a Charter-based challenge or indeed legislative reform for that reason. This is historically illiterate, as well as legally wrong. Justice Huscroft, fortunately, does not go anywhere near this line of reasoning.

The more significant mistake he does make, however, is the suggestion that “the Governor General invites the political party whose candidates win the most ridings to form a government”. That is constitutionally true if, but only if, the party in question has won an absolute majority of seats. Otherwise, as Philippe Lagassé, for example, has explained, and as I further discussed here, here, and here, as well as in a forthcoming article, the significance of a party winning more seats than any other is political, if any. Such a party is not constitutionally entitled to form a government, and another one, in coalition or indeed alone, may end up governing if it can secure the confidence of the House of Commons. This too is of no consequence for Justice Huscroft’s reasons, since the case has nothing to do with government formation, but quite a few people are mistaken about the relevant constitutional conventions, and it is unfortunate that they will have an obiter dictum in a Court of Appeal decision to support them in their views.

On, now, to happier things. Perhaps the most important of these is the way Justice Huscroft pushes back against the expansion of section 3 of the Charter, whose import, thanks to the Supreme Court, has long been unmoored from its text. On its face, section 3 provides Canadian citizens with the “the right to vote in an election of members of the House of Commons or of a legislative assembly and to be qualified for membership therein”. But the Supreme Court has long read it as protecting a far more nebulous purpose of ensuring “effective representation”. The Court of Appeal is, of course, bound by these decisions, but Justice Huscroft tries — pretty successfully, I think — to explain that they do not override constitutional text, and cannot be extended indefinitely. He points out that

Canadians did not adopt a generic charter of rights. We adopted a very specific charter of rights: the Canadian Charter of Rights and Freedoms, which establishes as supreme law the specific rights and freedoms it enumerates and so enshrines. The text of the Charter and the constitutional settlement it effects necessarily structure and delimit the scope of purposive interpretation. [46]

Even in cases where the Supreme Court took a broad approach to section, the right it protected was no more than

a right to participate in a democratic electoral process [which] does not entail a right to be represented by a candidate or party of one’s choice, or a right to have that candidate or party share in the political decision-making authority that flows from forming government. [53]

The first-past-the-post system does not interfere with the individual, participatory right enshrined in section 3. That it results in some candidates or parties being disadvantaged when individual votes are aggregated does not change this, and there is, accordingly, no Charter violation here.

The other claim against first-past-the-post was based on the violation of the Charter‘s equality provision, section 15(1), on the basis that first-past-the-post allegedly results in fewer women and members of minority groups being elected than a proportional representation system, resulting in unjustified adverse-impact discrimination. Having noted the considerable difficulties the Supreme Court has had in clarifying this area of the law, Justice Huscroft argues that

the first stage in the analysis under s. 15(1) – whether the legislation creates a distinction based on enumerated or analogous grounds, on its face or in its impact – depends on causation. If it were otherwise – if mere statistical disparity in the application of a law were sufficient at the first stage of the analysis – the scope of adverse impact discrimination would be so broad as to trivialize the concept. Given that all laws of general application apply imperfectly, findings of adverse impact discrimination would become routine. Discrimination would be all but impossible for legislators to avoid.

…

Identifying causation in the electoral system is especially difficult because electoral outcomes are multi-factorial: they flow not simply from the way in which ballots are translated into representation, but from myriad decisions made by political actors and parties, not only locally but also at the regional and national levels – decisions made long before millions of citizens decide how to cast their ballots. All of these decisions are made in the context of the SMP electoral system and the political incentives and disincentives it establishes. Correlation is obvious, but correlation is not causation. [71], [73]

Justice Huscroft then takes a bit of a detour to discuss an issue that I used to bring up when writing about the “empirical turn” cases, starting with this post on Canada (Attorney General) v Bedford, 2013 SCC 72. Here’s what I wrote:

I am very uncomfortable with the Supreme Court’s insistence on deference to the trial judge’s fact-finding, and the key role it played in the decision. Excessive reliance on and deference to first-instance fact-finding risks shifting power to people neither expected nor prepared to wield it, undermining what little accountability there is in the judicial review process, and making Charter litigation the tool of interest groups rather than of individual citizens harmed by the government.

And here is Justice Huscroft, having noted that the first instance judge found that the applicants did not establish the empirical support for their claim of discrimination and that under Bedford this finding is entitled to deference — and that the contrary finding would have been too:

Ultimately, Justice Huscroft agrees that there is no violation of equality rights, because the causal relationship between the electoral system and the lack of gender parity in election outcomes is not established:

No matter what electoral system is employed, over or underrepresentation of particular demographic groups may occur to a greater or lesser extent from time to time. Arguments based on over and underrepresentation lose sight of the fundamental fact of the matter: the people are entitled to vote for whomsoever they want, for any reason they want. This is the very essence of the freedom that lies at the heart of the right to vote. [90]

A word, finally, on the concurring opinion of Justice Dawe, who cautions that, contrary to what Justice Huscroft might seem to suggest, section 3 of the Charter “must constrain both the extent to which an electoral system can permissibly deviate from absolute voter parity, and the justifications that can permissibly be relied on by the legislature to support any such deviations” and that it is not the case “that every imaginable electoral system would be constitutionally compliant”. [102; emphasis in the original] I think that is right, if not quite for the reasons Justice Dawe suggests, which rely on a “right to effective representation” derived from the Supreme Court’s case law. As Justice Huscroft argues, that is not a right. But I don’t think it would be unfair to say that, past some (admittedly uncertain) threshold, the right to vote is rendered nugatory by very substantial departures from parity, or, conceivably, by other issues.

And I’m not sure that Justice Huscroft would actually disagree with that. He notes that the right to vote requires implementation — the voting process must be organized, the ballots must be counted, etc. There is, as this case shows, room for reasonable and wholly constitutional disagreement about how this is all to be done. But, even taking a limited view of the constitutional constraint the applies, it is not the case that any implementation programme, no matter how shoddy, ineffective, or indeed self-serving, would be immune from constitutional scrutiny.