Electoral Reform - the election-atlas.ca proposal

With the federal electoral reform committee finally up and running, I feel it's time to share my idea of what this country's electoral system should look like.

For those familiar with proportional representation systems around the world, this is essentially a regional open list system. It preserves proportionality of the vote on a regional basis, while allowing for local representation by candidates chosen by the voters instead of the party.


The country would be divided into somewhere between 50 and 60 districts. In the 10 provinces, they would return between 4 and 9 members each. The House of Commons would remain at 338 seats until the next census, with all provinces keeping their current number of MPs.

The territories, which only have enough population for 1 member each, would essentially remain first-past-the-post. Single-member districts may also be used in remote areas of the 10 provinces, such as Labrador or Northern Quebec.

Districts would be small enough that at least quasi-local representation would be ensured, while no quotas would be needed to keep fringe parties out.

Major redistributions every ten years would no longer be needed except in rare circumstances. The number of seats allocated to each region would simply be reapportioned based on population changes, similar to the current situation in Ireland. If a district became entitled to 10 or more seats, it would be split into two; likewise, if a district was only entitled to 3 or fewer, it would be merged into a larger one.

The Ballot

The ballot would look drastically different from what voters may be currently used to:

A sample ballot for southern and western New Brunswick, using candidates from the 2015 federal election. Using this ballot, voters could choose up to 5 candidates, regardless of party.

Electors will have as many votes on the ballot as members to be elected. This multi-member, multi-vote concept is already common for municipal councils across the country, mostly in smaller towns but also in bigger cities like Vancouver. Multi-member districts were also used for some provincial legislatures as late as the 1990s.

In the example above, voters in this district could choose up to 5 candidates in total, regardless of party. This allows them to split their vote between parties, if desired, without the need for separate systems (and separate "tiers" of MPs) like under a mixed-member proportional system.

It is expected major parties would continue to field a full slate of candidates in each district. Independents may run separately or in groups.

Counting the Votes

Seats would be allocated based on the combined number of votes for each party (or group of independents), using the D'Hondt method. Once the party standings are finalized, the candidates would be elected based on their own personal votes (for instance, if the Liberal party were to win 3 seats, the 3 Liberal candidates with the highest personal vote would be elected). Unlike in most PR systems, there would be no ranked list prepared by the party - the "order" would be completely up to the voters.

How the results for southern and western New Brunswick would be calculated under the D'Hondt method, using the actual results from the 2015 federal election. Note that the raw vote totals would be inflated by a factor of 5 under this proposal, since voters would have multiple selections.

In the example above, instead of the Liberals sweeping all 5 seats like they did under FPTP, the Conservatives would win 2, keeping effective local opposition. Had the NDP received 3000 more votes, they would too have won a seat at the expense of the Liberals; which brings up another point: a PR system would lessen the need for strategic voting.

Using geographical groupings of ridings across the country, the results of the 2015 election would look somewhat like this:

Party Pop. Vote PR Seats (%) FPTP Seats (%)
Liberal 39% 155 (46%) 184 (54%)
Conservative 32% 108 (32%) 99 (29%)
NDP 20% 60 (18%) 44 (13%)
Bloc 5% 13 (4%) 10 (3%)
Green 3% 2 (1%) 1 (0%)

The result would be closer to (but not truly) proportional as under FPTP. As voters would be more likely to vote for smaller parties (particularly the Greens) under a PR system, it is likely that their seat count would increase, and the "bonus" for the winning party would decrease.

Another formula (Sainte-Laguë) could be used to make results even more proportional on a national basis, but at the expense of proportionality on a local basis. For instance, in the New Brunswick example above, the NDP would win that fifth seat, but they would have half the seats of the Liberals on only 1/3 of their vote. I feel regional proportionality is more important than national - we are, after all, electing regional representatives.

Other Concerns

This open list system in particular also eliminates the quandary of voters liking a local candidate but not his or her party; as they could vote for that individual in particular but give the rest of their votes to another party.

The main drawback could be the loss of strictly local representation, particularly in rural areas. This can be rectified by ensuring parties nominate candidates from different areas of each region, and encouraging electors to vote for candidates from their local area.


I feel this proposal addresses most of the concerns about the current electoral system, as well as the issues raised so far by the electoral reform committee. Constituents would still have an MP from their local area (and more likely their party) to contact, the vote more closely matches the will of the people, there will be no need for two parallel electoral systems (and classes of MPs), northern and remote areas will still have proper representation, the concept of marking an X next to a candidate's name will still exist, and voters will still have a direct say on exactly which people represent them in Ottawa.

J.P. Kirby

July 2016