PDF
A Better Electoral System for Canada: Moderate Proportional Representation With Ranked Ballots
Jean-Pierre Derriennic

Canada’s current plurality vote system can create false majorities, lead to strategic voting and exacerbate regional cleavages, despite often bringing the stability of a coherent parliamentary majority government. Although proponents of reform may agree that the current system should be changed, they are often divided about what type of system should replace it. In this article, author Jean-Pierre Derriennic suggests two prominent reform models—a preferential/ranked ballot system and a moderate-form of proportional representation—could be combined to create a system that allows voters to cast ballots sincerely, reduces partisan regional polarization, and ensures stable coalition governments made up of parties that have broad popular appeal.

Electoral reform is needed in Canada to correct the major flaws in the voting system we have been using to date. It is not a majority system, as it is often called, but rather a plurality vote system, since a candidate can be elected with the support of less than half of the voters in that riding. Candidates simply need to get more than their opponents, which is sometimes called a “plurality”. This leads to the one positive aspect of this voting system, but also its main flaws.

The plurality vote system makes it easier to form coherent parliamentary majorities by granting, most often to the party that won the most support, a higher proportion of elected members than their share of the popular vote. This is the main argument made by proponents of this system. However, it can also allow false majorities to form, when one party gets more elected members than a rival party that won more of the popular vote. It exaggerates the conflicts that exist between the different regions of Canada, by preventing either the government majority or the opposition from having any representation in certain provinces. Quite often it forces voters to vote strategically rather than sincerely, and creates enormous disparities in the political influence enjoyed by people depending on the number of electors in their riding, and especially between ridings in which the gaps between candidates are narrow and those in which the gaps are quite wide.

It is possible to rectify those flaws while preserving the only advantage of the current system, that is, the possibility of forming coherent parliamentary majorities. To do so, two methods must be used: moderate proportional representation and ranked ballots. Ranked ballots can be applied in single-member ridings, which is what we have now, or in ridings electing several members proportionately among the parties. These two methods are therefore not mutually exclusive, as I explained in a short book published recently by Les Presses de l’Université Laval entitled Un meilleur système électoral pour le Canada / A Better Electoral System for Canada.

Ranked ballots are highly recommended because it puts voters in a much better moral and intellectual position than the current single-choice voting system, which often forces people to vote strategically: they often have to choose between voting sincerely, for their preferred candidate or party, and voting effectively, for the least detestable of those who have any chance of winning. With the possibility of ranking candidates in order of preference, votes for the candidates with the fewest first preferences are transferred, as the results are calculated, to other candidates based on subsequent preferences indicated by voters. People can express a sincere first preference, even if it goes to a candidate who has no chance of winning, then rank the other candidates in order of preference, with voters’ least favourite candidate ranked last. Sincere voting is effective, whereas strategic voting is almost never useful.

The information needed to vote strategically is not easily accessible to voters, which makes them more susceptible to manipulation by polls and rumours. There is a moral equivalence between secret ballot voting and ranked ballots. Secret ballot voting protects voters from undue influence. That is why we believe that, in a democracy, voting must be secret. Similarly, ranked ballots protect voters from being manipulated, and it should therefore always be considered a deontological rule.

When applied in single-member ridings, ranked ballots can produce stronger parliamentary majorities than the plurality vote system, but those majorities are more authentic, because all members are elected with a majority in their riding. Ranked ballots give one party an advantage over another only if a candidate is a Condorcet winner; that is, one who would have won over all of his/her opponents in separate elections against each of them. The Condorcet winner is not necessarily the candidate who has the most committed supporters, but rather the one who is most acceptable to the highest number of voters. Ranked ballots reduce the likelihood of a party winning a parliamentary majority without having a Condorcet winner.

Ranked ballots do not favour any one party or ideology; it means that elected officials must win a majority in each riding, and favours moderate parties capable of winning seats thanks to the second preferences of other parties’ voters. This added bonus of moderation is desirable in a democracy, because it incites political parties to avoid the simplistic arguments that emphasize the differences between them, and not always benefits the same party. In 2015, ranked ballots would probably have given an advantage to the Liberal Party. In 1993, the Conservative Party won two seats with 16.4 per cent of the vote, while the Reform Party won 52 seats with 18.69 per cent of the vote. With ranked ballots, the Progressive Conservative Party, which was more moderate, would probably have gotten more second preferences than the Reform Party, which was more radical. It also would have won in terms of the number of votes, and perhaps even in terms of the number of elected members.

Ranked ballots would not give smaller parties more elected members, but they would very likely win more of the popular vote. The Green Party, for example, obtained only 3.4 per cent of the vote in 2015, which is primarily a result of the fact that, almost everywhere, a vote for the Greens was thought to be a wasted vote. With ranked ballots, the fear of wasting a vote would vanish, and the Greens would likely win a greater popular vote. They might not win any more seats, but they would have better political visibility and a good starting point from which to further develop. More importantly, the other parties would know that some of their members were elected thanks to the transfer of the second and third preferences of the Green Party’s voters and might, therefore, be more inclined to take their concerns into account.

When applied in single-member ridings, a ranked ballot is one reform that would be quite easy to put in place quickly because there would be no need to change the number or the borders of the existing ridings. This voting system would have significant benefits compared to the plurality vote system we have been using for so long. It would allow voters to vote sincerely and effectively without having to resort to strategic voting. It would help create a party system adapted to the proper functioning of our democracy, with moderate, large parties, while small parties that find it harder to survive would have more known popular support, and large parties would be forced to pay attention to the small parties’ voters. It would be even better if this voting system were both preferential and proportional.

If proportional representation is chosen, two mistakes must be avoided: creating ridings that elect a very unequal number of members, and creating ridings that elect a large number of members. Unequal ridings are unfair because the choice offered to voters varies depending on where they vote. When ridings elect a large number of members, it favours the proliferation of political parties; this could lead to assemblies that are unable to make any decisions. We should create ridings that each elect a small number of members, in order to achieve what Vincent Lemieux called “moderate proportional representation”.

In Canada, electoral boundaries must take the provinces into account. The smallest province, Prince Edward Island, has four MPs and could form one riding with four seats. To limit inequalities among the ridings, the other provinces would have to be divided up into ridings of three, four or five seats. To limit the proliferation of political parties as much as possible, creating ridings of three or occasionally four seats would be the most advisable option. The new electoral map could be created with the following criteria in mind: without changing current electoral boundaries, three of them could be grouped together, or occasionally four, when four are needed to respect the number of MPs by province. As the new groupings are being created, every effort should be made to ensure the least variation possible in the ratio of population per member in the new ridings thereby created. Canada has very knowledgeable experts on electoral systems who would be able to take care of that in just a few days.

This reform would be marginally more complicated than any reform that preserves the current single-member ridings, but it could definitely be implemented fairly quickly. This would be much simpler than creating a mixed system that includes single-member ridings and additional members elected in a proportional, compensatory or parallel system. In order to create a mixed system, we would either need to increase the number of seats in the House of Commons considerably, or decrease the number of single-member ridings, in other words, completely redraw the electoral map, which would be very time consuming and would raise a number of concerns and political protests. A mixed system would involve a complicated reform, and the only advantage would be to possibly yield a proportional result among the parties while preserving single-member ridings. In three-seat ridings in a moderate proportional representation system, MPs would be just as accessible to their constituents as they are in single-member ridings, and many Canadians would no doubt appreciate having access to several MPs when they need to reach out for help with something. Moreover, they could reach out to either a member of the government majority or to a member of the opposition.

With three- or four-seat ridings in a proportional representation system, the number of political parties represented in the House of Commons would likely stay the same as today, but they would each have a ratio of MPs roughly proportional to the actual public support they received, and their geographic distribution would be very different. It is very unlikely that one party would win all three seats in a given riding, and completely impossible that one party would win all the seats in a province. This would eliminate some of the extreme differences among the various regions of the country, which is one of the most harmful aspects of a single-member plurality vote system for Canadian politics.

Moderate proportional representation would yield more equitable results among the major parties and help the parliamentary system run more smoothly, but it would still be very hard on small parties. To elect an MP in a four-seat riding, a party would have to win about 20 per cent of the vote, and in a three-seat riding, about 25 per cent. That is why it is highly recommended that voters be allowed to rank their preferences among the various parties. Votes for small parties would therefore not be wasted, because voters’ subsequent preferences would be taken into account. Those voters may not be represented by their preferred candidates, but MPs would be more inclined to take their concerns into account, because they would have been elected thanks to those second or third preferences.

A ranked ballot is always advisable because it allows voters to vote sincerely without any fear of wasting their vote. It is a good complement to moderate proportional representation, which makes it harder for small parties to survive. It is also recommended because of its impact on larger parties, which are more inclined to avoid exaggerating the ideological differences among them in order to increase their chances of winning the subsequent votes of their rivals’ voters, as they will need those votes to win seats in certain ridings. Lastly—and this is perhaps the main advantage of a ranked ballot for proportional representation—it allows voters to influence the formation of coalition governments.

Moderate proportional representation will not result in a proliferation of political parties, which would create irreparable political instability. However, it would make it harder for one political party to win a parliamentary majority, and coalition governments would therefore need to be formed more often. Coalition governments exist in some of the best-governed countries in the world today, while in other countries, they are a source of instability and decision paralysis. The first type of result can be seen in countries where alliances among the political parties are ideologically aligned and accepted by those who voted for them. The second type of result exists in countries where ideological alliances are incompatible and shocking to voters. Knowing which party won most of the second preference votes of each party’s voters after an election gives a very clear indication of which coalitions are acceptable and which are not.

In closing, if electoral reform results in ranked ballots in single-member ridings, this would be a vast improvement over our current electoral system. A proportional system with three- or four-seat ridings would be very good. Adding ranked ballots would be an excellent reform. It would be very unfortunate if the debate on electoral reform were to turn into a division between advocates of the preferential system and advocates of the proportional system, when the two systems could easily be combined and complement one another very nicely.