Balancing Trade-offs for a Win-Win Solution
Electoral reformers in Canada have espoused two main ways of increasing the proportionality of our electoral system:
- the use of top-up seats to compensate for the disproportionality of single member ridings (the MMP approach), or
- multi-member ridings to allow several MPs representing different points of view to be elected in each region (the STV approach).
While MMP and STV both deliver proportional results, some critics have expressed concern about the increased size of ridings that result when using these approaches in rural settings. Under MMP, single-member ridings would remain, but would typically be about 60% larger. Under STV, the smallest multi-member riding that could be envisaged in rural areas would involve pairing two ridings. Leaving some single-member ridings as-is would be an option, but that would do nothing for proportionality or voter equality in those ridings.
This brings us to recommend consideration of a hybrid model that would combine the use of multi-member ridings and top-up seats to meet the different needs of both rural and urban areas, while protecting the objective of proportionality. We call this hybrid Rural-Urban Proportional Representation (RU-PR). Rural-Urban PR uses single-member ridings as may be deemed appropriate in rural and “small-urban” areas involving cities and towns of less than 100,000 residents. In large urban and surrounding areas, it uses multi-member ridings. Proportionality is maintained for both urban and rural voters by adding a limited number of regional top-up seats in an MMP-like fashion.
The advantages of RU-PR are that it makes it unnecessary to create large multi-member ridings in the search for proportionality. Areas in or near large urban centres, where 60% of Canadians live, could have multi-member ridings of modest size as in Ireland or Northern Ireland (where ridings range from 3-6 MPs). Rural areas and small-urban areas could use single-member or dual-member ridings as appropriate. Relatively few top-up seats would be required to compensate for any remaining disproportionalities because the first-round results would already be quite proportional.
Figure 1 shows RU-PR’s relationship to other models. STV (top left) uses multi-member ridings and no top-up seats. MMP (bottom right) uses many top-up seats and single-member ridings. RU-PR occupies a middle ground. It uses a mix of single-member and multi-member ridings of modest size, using a small number of regional top-up seats to enhance the proportionality of the system.
The Rural-Urban PR model draws inspiration from several sources. It builds on a suggestion proposed by Canada’s former Chief Electoral Officer, Jean-Pierre Kingsley, that we use multi-member ridings of three to five MPs in urban areas while leaving rural and remote areas as single-member ridings. However, it adds a greater element of proportionality to that proposal.
RU-PR draws further inspiration from Sweden’s electoral system. After electing most MPs in small multi-member ridings using a list-PR voting system, Sweden reserves 11 percent of MP seats to be allocated on a top-up basis, country-wide, to compensate for disproportionalities at the local level. Because only a few top-up MPs are needed, Sweden uses the simple one-ballot “best runners-up” method to elect them. Canada could do the same.
Figure 2 illustrates how single-member ridings, multi-member ridings, and top-up seats combine to give a more proportional result. In this figure, the gold oval represents a region containing two single member ridings and three multi-member ridings. The red oval shows that this region also has two top-up seats for a total of 15 MPs.
In Canada, we estimate the RU-PR hybrid model would require a top-up layer in the order of only 13% to 15% of total seats, much fewer than the 35-40% usually envisaged under MMP.
RU-PR can be tailored to Canada’s specific needs by adjusting a number of design features, including the size of ridings, the number of single-member ridings, how MPs are elected and how top-up seats are filled. These are discussed in more detail in the “Design choices” section of this appendix.
Figures 3 and 4, below, illustrate two examples of how RU-PR might be configured for the province of Alberta. Both examples show a mix of single-member ridings (gold), multi-member ridings (green), and regional top-up seats (red). The examples are based on the same 34 ridings as now but with an additional six or four top-up seats, respectively. The number of top-up seats varies because it depends on the degree of proportionality already achieved through the use of multi-member ridings.
The two examples shown increase the size of the House by adding the top-up seats to the current 338 seats. This allows existing single-member ridings to stay the same and multi-member ridings to be grouped together without the need for redistricting in most cases.
An alternative maintains the current size of the House by enlarging regular ridings slightly so that fewer MPs are required. The seats thus freed up in the House become the top-up seats. This is the typical approach when implementing MMP. However, RU-PR needs a riding size increase of only 13-20%, which is far less than the 60% typically required by MMP.
In both examples, the urban ridings in Calgary and Edmonton are grouped into multi-member ridings of 5-6 MPs each. However, the 13 smaller rural and small-urban ridings are treated differently in the two examples: in one, the choice is to keep most of them as single-member ridings; in the other, the choice is to use small multi-member ridings of two or three MPs each wherever possible. Middle-ground options could also be envisaged, but we will limit ourselves to these two examples for illustrative purposes.
In smaller provinces, the top-up seats would be allocated province-wide. In larger provinces, an average of three top-up seats would normally be needed per region of about 20 MPs. Although the 13-15% of MPs elected to fill top-up seats would represent regions covering a number of multi-member and single-member ridings, their role is to provide a voice for voters who were unable to secure representation at the riding level.
For example, one might be a Green Party MP in northern Alberta. Such an MP would represent the voice of Green Party voters throughout northern Alberta and might hold office hours in several parts of the region, similar to regional MPs in Scotland. Of course that single MP would not be the sole MP representing the ridings of that region. Supporters of any party, including the Green Party, would have the choice of calling on their single-member MP if they have one, or any of the multi-member MPs in their riding.
A Kingsley-inspired example
The first example that we would like to put forward to illustrate the Rural-Urban PR concept is one inspired by Jean-Pierre Kingsley’s proposal alluded to earlier. Country-wide, we estimate that this option might retain about 25% of total seats as single-member ridings. This would make it easier for these MPs and their constituents to maintain a link with each other in rural and small-urban ridings. In this Alberta example, we have kept 11 rural and small urban ridings as single-member ridings. Red Deer might have 2 MPs. In total, this configuration would have 11 single-member ridings, 23 seats in multi-member ridings and six top-up seats, for a total of 40, under the additional seats approach discussed earlier. The six top-up seats could be divided between two regions – one comprising the top two-thirds of the province and one in the southern part of the province.
Example emphasizing multi-member ridings
Alternatively, we can consider what the model would look like if we were to minimize the use of single-member ridings by using two- or three-member ridings rather than single-member ridings in most rural areas. This would enable more rural voters to have a local representative from their chosen party.
Figure 4 retains only two seats as single-member ridings for illustrative purposes. The dual-member riding shown at the north-west corner of the map is an interesting case, because it illustrates that the case for single-member ridings does not depend only upon the remoteness or size of ridings. Based on the distribution of the population in this case, the two current ridings of Peace River—Westlock and Grande Prairie—Mackenzie would seem just as easy to manage as a two-seater, in particular the area around Peace River, which is currently split in half by the two ridings. These are the sort of considerations that would be taken into account to determine which ridings to leave as single-member ridings and how many there should be to ensure a proper balance between manageability and proportionality.
Because of the proportionality gains achieved by minimizing the number of single-seat ridings in this model, there would be less need for top-up seats. This example calls for only four top-up seats, split between two regions. This example illustrates what we had in mind in Appendix 11 when discussing the STV+ option.
The above examples illustrate the flexibility of the Rural-Urban PR hybrid that we are proposing for consideration. This flexibility allows the RU-PR model to be tailored to varying circumstances across Canada.
Size of Parliament
RU-PR requires only a small number of top-up seats compared to MMP. There are two choices when making room for these top-up seats:
- Top-up seats could be in addition to the existing 338 existing seats in Parliament. This would increase the size of parliament to about 380-390 MPs. Any single-member ridings would stay the same, and multi-member ridings could be created by simply combining existing ridings with no other changes to boundaries in most cases.
- If the choice is made to keep Parliament the same size, the Electoral Boundaries Commissions in each province could be asked to slightly increase the size of ridings to make room for the top-up seats.
Balance of single-member and multi-member ridings
The most obvious design choice concerns the balance to be achieved between the number of single-member and multi-member ridings. Rural-Urban PR would achieve a significant level of proportionality by using multi-member ridings in large urban and surrounding areas. Less densely populated ridings could then be either single- or multi-member. The more single-member ridings are used, the more top-up seats may be needed to achieve a satisfactory level of proportionality.
Electing riding MPs
MPs in multi-member ridings could be elected using some form of list-PR or STV (see Appendix 11 for a discussion of the advantages of using STV). MPs in single-member ridings could be elected with preferential ballots or FPTP. The ballots used in both the single- and multi-member ridings could be similar (ranked ballots everywhere or regular ballots using a simple X everywhere).
Filling top-up seats
Top-up seats could be filled using either an open list system or a best runners-up system.
Use of a best runners-up (no-list) system makes it unnecessary to have a separate top-up list and avoids the need for a second ballot. Voters indicate their choice for local MP, either with a single X as in Sweden or with a ranked ballot. Top-up seats are filled by determining which parties win the top-up seats in each region (following the same approach as MMP) and then awarding the seats to the strongest remaining candidate from the most underrepresented district for each party, just as Sweden does it.
Comparing Electoral Options for Canada
The following table summarizes the key features of various electoral systems. All of the proportional systems put forward in this submission yield medium to high levels of proportionality. What changes with Rural-Urban PR is that one can achieve the same high level of proportionality with relatively small multi-member ridings by adding a limited number of top-up seats.
|RU-PR Example 1||RU-PR Example 2|
|MPs from single-member ridings||100%||100%||60-65%||Excep- tionally1||5-25%||20-25%||5%|
|MPs from multi-member ridings||0%||0%||0%||95-
|MPs from regional top-up seats||0%||0%||35-40%||0%||13-15%||15%||About 15%|
|Size of top-up regions||N/A||N/A||6-19, average 8-14 MPs||N/A||7-20 MPs||About 20 MPs||7-20 MPs|
|Degree of proportionality||Low||Low2||Medium to High3||Medium to High4||High to Very High||High to Very High5||High to Very High|
- STV normally does not have single-member ridings, but could if local conditions make it difficult to combine with other ridings.
- Better reflects voter preferences locally but could be even more disproportionate than FPTP in Parliament.
- Increases with size of region and % of top-up seats.
- Medium level of proportionality with fairly small regions, high proportionality with larger regions.
- The large number of single-member ridings reduces proportionality. This can be compensated for by having larger top-up regions, more top-up seats, and more or larger multi-member ridings.
In summary, these Rural-Urban PR models feature:
- single-member or 2-3-member ridings in areas where population density is low,
- the benefits of multi-member ridings where population density is higher,
- a top-up share of about 15 percent of total seats, achieved by increasing the size of the House or the size of each riding by 15 percent, a minor change.
Different applications use different approaches to give each voter an effective voice. Features of the model can be adjusted to good effect in each region of the country to provide a made-in-Canada solution that provides the desired level of proportionality while still managing differences in riding sizes between urban and rural areas and remaining sensitive to local concerns and preferences.