Appendix 10: Made-in-Canada MMP

A widely-supported option for Canada

Internationally, MMP has been used in Germany since the mid-1940s, in New Zealand since 1994, and for parliamentary elections in Scotland and Wales since 1999. In 1998, the Jenkins Commission recommended a moderately proportional form of MMP for the UK that featured ranked ballots for local seats.

No PR option for Canada has been more often put forward than MMP. Since 1977, 13 electoral reform processes have looked at options for Canada, and ten of these have proposed some form of MMP. The first recommendations to point towards MMP as an option for Canada were the Pépin-Robarts Commission (1979), the Estates General Commission in Québec (2003), and the Law Commission of Canada (2004).

We note from the outset that Fair Vote Canada supports only models with no closed lists.

The extensive process used by the Law Commission lasted for three years, beginning in 2001. It Included 15 public hearings, the use of an internet questionnaire, and more than 30 other public meetings. The 229 page Law Commission of Canada report titled “Voting Counts: Electoral Reform for Canada” has become a key reference document on PR at the federal level. It provides a good statement on the rationale for such a system in Canada and includes a number of recommendations worth reproducing. Here are some highlights from the Executive Summary.   

…the Commission’s goal was to balance the benefits of introducing some element of proportionality into the existing system with the capacity to maintain accountable government, most notably as a direct link between elected politicians and their constituents. The Report, therefore, examines alternative systems from the premise that constituencies should stay small enough to maintain the Member of Parliament–constituent relationship. After robust consultations, the Report also accepted the premise that there is little appetite for substantially increasing the size of the House of Commons to accommodate a new electoral system. Finally, the report is based on the premise that changes to the electoral system should be made without a process of constitutional amendment.

Using an approach similar to that proposed in the motion to create the ERRE committee, the Commission chose 10 criteria for assessing eight electoral systems options for Canada. It is worth citing those criteria for the sake of comparison:

The Commission’s main recommendation was that adding an element of proportionality, as inspired by the MMP system currently used in Scotland, would be the most appropriate path forward for Canada. MMP retains a majority of local MPs, but reserves a certain number of seats to be used as top-ups on a regional basis to compensate for the disproportionalities of FPTP. The aim is for the total number of MPs in a region to match the vote share of each party in that region.

The report cites a number of potential benefits:

The Commission’s overall recommendation was to add “an element of proportionality” to Canada’s electoral system and adopt an MMP electoral system.

The basic features of the model proposed by the Law Commission are found in recommendations 3 to 5:

In short, with two votes, it’s both personal and proportional.

There followed several recommendations (6 through 10, pp. 112 to 116) to promote representation by a diversity of voices in parliament (women, minority groups and youth).

They also suggested that an ad hoc Parliamentary committee should review the new electoral system after three general elections have been conducted under the new electoral rules.  

Nine months after the Law Commission of Canada Report, the New Brunswick Commission on Legislative Democracy recommended a similar model, with regions of 14 MLAs, nine local and five regional. Premier Bernard Lord said he would proceed, but then lost power in an early election.

Seven other reports since 2003 have all recommended a mixed member model for a province.

Open Lists

A feature that must be considered when designing an MMP system is how to select who will represent each party in the regional seats. Under a closed list model (which FVC does not propose), the parties present ranked lists of candidates, and voters vote for the list. If two regional MPs are elected for a given party, the winners would be the two candidates at the top of the list.

In an open list system, it is voters themselves who determine the order of candidates, by voting for their personal choice of candidate for a regional seat. Under a flexible list system – the recommendation of the Law Commission – voters have a choice of voting for the candidate or for the party list as a whole. Candidates that receive a certain quota of the party’s vote (8% in the Law Commission’s proposal) are bumped to the top of the list.

An important feature of the Law Commission’s made-in-Canada model, was its recommendation 5, which proposed the use of “flexible” lists. As they explained,

Based on the feedback received during our consultation process, many Canadian voters would also most likely desire the flexibility of open lists in an MMP system. In essence, allowing voters to choose a candidate from the list provides voters with the ability to select a specific individual and hold them accountable for their actions should they be elected.

The flexible open list method was also recommended by the Jenkins Commission in the UK. Their colourful explanation accurately predicted why closed lists would be rejected in Canada. As they put it, additional members locally anchored are, 

more easily assimilable [sic] into the political culture and indeed the Parliamentary system than would be a flock of unattached birds clouding the sky and wheeling under central party directions.

When Sweden changed from closed list to flexible list, some women feared male backlash would result in women being moved down the list. As it turned out, more women were moved up than down. Similarly, many Canadian voters might jump at the chance to vote for a female regional candidate, especially if their party had nominated a man for the local seat.

Fair Vote Canada shares the Law Commission’s view that open or flexible lists are likely to be more acceptable to Canadians than closed lists. There is evidence that the closed lists included in the referendum proposals in Ontario and PEI played into the failure of those referendums.

We call the MMP option with open or flexible lists “Personal MMP” to distinguish it from what was proposed in the Ontario referendum. The Law Commission proposal is one example of an open or flexible list model. Alternatively, the German province of Bavaria has a fully open model well worth considering. In Bavaria, one’s second vote must be for a regional candidate. The ranking of the regional candidates is by the voters alone, so all MPs are directly accountable to the electorate.  

A frequent argument in favour of closed lists is that it allows the parties to elect a more diverse roster of candidates, in terms of gender, for example. However, flexible or open lists may provide similar benefits, provided that the party offers a diverse range of candidates, as it would naturally want to do. For instance, polls show that 90 percent of Canadian voters want to see more women elected. If they can choose from several of their party’s regional candidates, they’ll almost certainly elect more women.

The Appeal of MMP as an Electoral Reform Option for Canada

As a proportional system, MMP offers all of the advantages of PR as discussed in the main part of this submission. We will not repeat those advantages here. However, MMP may appeal to committee members for reasons that are specific to this model, presented here in terms of the five ERRE principles, as we did in the main text of this submission.

Effectiveness and legitimacy

MMP is a proportional system. It is “mixed” because it uses both a winner-take-all approach for local seats combined with a party-list system for the compensatory regional seats, but if it is well designed, MMP can achieve a relatively high level of proportionality. The key is to ensure that the top-up regions are of sufficient size (perhaps 14 MPs per region on average, or at least eight) and that the share of top-up seats is sufficient (from 36% to 45% of the total).   

As detailed below, based on how Canadians voted in 2015, voters for all three parties would have elected local or regional MPs in almost all of the 42 regions across Canada under an MMP model with 8-MP regions.


With two votes, you can vote for the party you want in government. And you can also vote for the local candidate you like best regardless of party, without hurting your party, since it’s the second (regional) ballot that determines the party make-up of the legislature. About 32% of voters split their ballots this way in New Zealand’s MMP system. Increased voter choice of this sort and making every vote count due to the proportionality of the system are likely to make voting more appealing.

High performing MPs and star candidates may be able to attract support from voters of different persuasions, who can support them without sacrificing the interests of the party they also support. Local MPs can therefore become more independent.   

Accessibility and inclusiveness

An MMP ballot is only slightly more complicated than a FPTP ballot because voters have two votes, one for a local candidate, and one for the party or regional party candidate of their choice. However, there is a tradeoff between the simplicity of the ballot and direct accountability to the voter. When the Ontario Citizens’ Assembly submitted its proposal in 2007, they wanted the system to be as simple as possible, so they opted for a closed-list model with a single list for the whole province. This is how the system works in Germany’s federal elections. It is very simple: you cast one vote for a local candidate, and one for a party’s closed list. But this model did not resonate well with Ontario voters, and we do not propose it.

With the variation used in the German province of Bavaria, you vote for your preferred regional candidate of your party, as the Law Commission of Canada recommended. Every candidate faces the voters directly, but in return, you have a bigger ballot. Fair Vote Canada considers that a bigger ballot would be a small price to pay for increased accountability to the voter.


Public trust in the election process will be enhanced when false majorities no longer occur, and when rural and urban voters in every region have effective votes and fair representation in both government and opposition. To quote the Jenkins Commission in the UK on “safe seats,” ”many voters pass their entire adult lives without any realistic hope of influencing a result. In these circumstances it is perhaps remarkable that general election turnouts remain at a respectable level.”

MMP does not pose any particular challenges to the integrity of the voter system. The regional result will be certain when each riding in the region has reported.

Local representation

The whole point of MMP is to retain a strong element of local representation of the kind to which Canadians are accustomed. Local ridings will necessarily be somewhat larger, to make room for the top-up seats, but in return, voters will have more than one MP that they can turn to for support or with whom to discuss policy issues.

Voters are guaranteed two things which in combination represent better local representation:

  1. A local MP who will champion their area, accountable only to local voters.
  2. An MP whose views best reflect their preferences, whom they helped to elect and with whom they may have greater affinity.

Proponents say this gives voters the best of both worlds.  

Some people wonder how regional MPs would be able to service a region that is several times the size of existing ridings. Most will serve constituents in their part of the region, although some may also choose to focus more on regional or national issues. Our simulations show that in 2015, an MMP model in Saskatchewan would have elected two Liberal MPs at the regional level. These MPs might be based in Saskatoon, Prince Albert, or Regina, but they would likely have additional offices in North Battleford, Yorkton, or Swift Current, and perhaps elsewhere, just as Conservative MP Robert Kitchen has offices in Estevan, Weyburn and Moosomin. This is how it’s done in Scotland, where two regional MPs from a party will normally split the region between them for constituency service purposes, and hold office hours rotating across their region or their part of it.    

Design Issues: Trade-offs and Options

Should Canada opt for MMP as a way to retain local MPs in the way to which Canadians have become accustomed, this still leaves a number of parameters to be defined. We consider some of those below.

Region size and share of top-up seats

The proportionality of MMP is determined by the size of the MMP regions and the share of top-up seats in each region. The size of the region matters, because it defines the threshold required to win a compensatory seat at the regional level. The larger the region, the easier it will be to ensure that the preferences of all voters can be accommodated. In a 15 member region, the threshold to be sure of winning a seat is 6.25% (1/16). In an 8 member region, it is 11%, although the “highest remainder” calculation often lets a smaller percent elect an MP.

The number of top-up seats matters because a minimum number are likely to be needed to compensate distortions in proportionality in the distribution of local seats. So if one party sweeps all of the local seats in a region, as often happens, one needs enough top-up seats to ensure proportionality. However, the larger the region and the larger the share of top-up seats, the less voters are going to feel personally connected to their MPs. This suggests a need to find the right balance.

Many experts recommend 40% regional MPs, where possible, with variations depending on geographical factors. For instance, in Northern Ontario, with nine MPs and ridings that are already very large, it would be difficult to envisage more than 33% regional MPs. In urban areas, the number of regional MPs could be larger to ensure a higher level of proportionality.

Scotland uses regions of 16 MPs (9 local MPs, 7 regional MPs). Wales uses regions of 12 MPs (8 local MPs, 4 regional MPs). In Germany, the “region” is the whole province and the share of regional MPs is 50%, so the system is highly proportional.

The Law Commission had only about 35 percent top-up MPs, to avoid having to increase the size of local ridings too much.1

The Jenkins Committee in the UK recommended a moderately proportionate system, with local regions averaging only eight MPs. This helps keep the regional MPs closer to their electorate, but makes the system less proportional.

The Hon. Stéphane Dion has suggested that the configuration of PR regions in different provinces might divide Canada into different political microclimates. To address this concern, we need to seek some uniformity in the size of MMP regions across the country. Currently, four provinces have ten to 14 MPs, and four have considerably more. Only PEI and Newfoundland, with four and seven MPs respectively, have less than ten; the three Northern territories have one each. If we were to propose an average of 11 or 12 MPs per region, including in the four larger provinces, we would have about 30 or 32 MMP regions across Canada, all with the same degree of proportionality. A new Boundaries Commission in each province would set the regional boundaries, using the parameters set by Parliament.

Parliament might decide that the average region should have 14 MPs, eight MPs, or some other number. With an eight-MP average, region size could range from six to 11 (four in PEI.) The exact regional boundaries would be set by new Boundaries Commissions in each province, within the parameters set by Parliament. (See sample regions below.)

Allowance for dual candidacies

Normally every mixed compensatory system lets parties sometimes nominate the same person for both local MP and regional MP, rather than asking a candidate to gamble on which position to run for. Each party would hold regional nomination meetings and/or vote online to nominate their regional candidates. These would often be the same people nominated locally, plus a few additional regional candidates. The meeting would decide what rank order each would have on the regional ballot, although with flexible or open lists, it is voters in the region who would have the final choice. Any local candidates who win local seats are of course removed from the count for regional seats.

By-election rules

In case of a resignation or death of a regional MP during a term, the party’s runner-up moves up into the seat. No working MMP model has by-elections for regional MPs. As the Jenkins Commission pointed out, if a region-wide contest were to take place “it would almost by definition result in the victory of the predominant party in the area, thus negating the essential purpose of the Top-up seats.


The Law Commission recommended that the right to nominate candidates for regional top-up seats should be limited to those parties which have candidates standing for election in at least one-third of the ridings within the province. The Jenkins Commission  recommended an even higher requirement of at least 50 percent. The aim is to prevent a possible distortion of the system by parties pretending to split into twin decoy parties for the regional seats, the trick which Berlusconi invented to unhinge the compensatory mechanism, and sabotage Italy’s voting system by de facto establishing a parallel system.

Ranked ballots or FPTP?

One way to increase voter choice for the local seats would be to introduce the use of ranked ballots, thus reducing any perceived need for strategic voting that might remain or the risk of vote-splitting. Whether this would make a significant difference is difficult to say. In a few regions, it might generate a local sweep by one party, making the number of compensatory MPs inadequate. In a few others, it might prevent such a sweep. The Jenkins Commission warned that “on its own, the effects of the Alternative Vote are disturbingly unpredictable.” However, the introduction of ranked ballots may not do significant harm so long as the region size and number of top-up seats are large enough to compensate for any disproportionalities that might emerge.


PR systems often include a legal threshold of approximately four or five percent before a party is allowed to win a top-up seat. The intention is to limit the number of parties in parliament to a manageable number. In a made-in-Canada MMP system, the threshold would apply to the regional top-up seats on a province-by-province basis. However, with region sizes from six to 15 MPs, a party would need over 5 percent to win a regional seat anyway. The Law Commission recommended no legal threshold. However, a legal threshold would ensure that no regional micro-party might elect a single MP.

Rounding formula

Since the number of party seats corresponding to each party is unlikely ever to be a round number, some sort of rounding formula is required to determine the winners. “Highest remainder” is the simplest and most transparent. Thus, if Party A deserves 3.3 MPs, Party B deserves 2.2, and Party C deserves 1.5, in a seven-MP district, the seventh seat will go to Party C. which enjoys a remainder of .5 compared to .3 and .2 for the other two parties.

While this method seems very simple and intuitive, more complicated methods do exist, including the D’Hondt or “highest average” method used in Scotland. The choice of method matters, because the D’Hondt method tends to favour large parties. In Scotland it was a factor in giving the Scottish National Party a majority government with only 44% of the vote in 2011.   

Regional preferential ballot

You could use a preferential ballot to let voters rank all the regional candidates of the party they wish to support with their second ballot. This could be valuable to help voters make their vote for a regional candidate count even if their first preference is for a regional candidate who ends up winning locally. This would increase voter choice, but would complicate both that choice and the counting process, so there is a trade-off involved.

The best runner-up or no-list MMP option

One way to simplify balloting for the top-up seats is to allocate top-ups using a best runners up approach. Under this model, used in Germany’s Baden-Wurttemberg province, the top-up seats are allocated by drawing from a party’s defeated candidates in the region, starting with the candidate who got the most support without being elected. This means that some ridings will end up with more than one MP.

The Ontario Legislature’s Select Committee was interested enough in this model in 2005 that their executive travelled to Stuttgart to study it. One of their findings was that, unlike in most German elections, Baden-Wurttemberg has elected no more women than Ontario or Canada, because all candidates are nominated one at a time.

Exceptional ridings

There are likely to exist some ridings whose size cannot be increased for exceptional reasons in the northern part of the country and Labrador.

Sample regions

The size of top-up regions will vary depending upon the population of different provinces. In PEI, it cannot be anything but four. Elsewhere, it could range from six MPs to 18. The aim in defining these top-up regions should be to ensure that all MPs are accountable to real communities, or as the Jenkins Commission put it, locally anchored to small areas. With smaller regions, Scarborough voters would not be represented by an MP from Etobicoke, and Kingston voters would not be represented by an MP from Ottawa.

Below are examples of how MMP regions could be configured, with an average size of eight MPs per region. This yields 42 moderately proportional regions across the country.

Ontario (15 regions)

Quebec (10 regions)

Manitoba (2 regions)

Saskatchewan (2 regions)

Alberta (4 regions)

British Columbia (5 regions)

In Atlantic Canada, the “top-up region” is the whole province.


According to simulations of how the 2015 election results would have been different under MMP with 8-MP regions and a 38% share of regional MPs, the result would be quite highly proportional, though not perfectly so.

The limitations on the model are two: a) with very strong regional sweeps by one party such as what was achieved by the Liberals in the Atlantic provinces and in the Toronto area, it may not be possible to fully compensate for the resulting disproportionalities without a larger number of top-up seats; b) with the relatively high share of the vote needed to secure a seat in moderately-sized regions, the Green Party would not be able to obtain a share of seats corresponding to its 2015 share of the vote.

Here are the minor exceptions to proportionality emerging in our simulations:

Significantly, our simulations show that all three of the larger parties would have elected local or regional MPs in each of the 42 regions across Canada with minor exceptions (no Conservatives in Montreal-est region or the Outaouais—Abitibi-Témiscamingue—Nord region). Bloc voters would have elected MPs in each of Quebec’s regions except Montreal West region.

This is a significant finding, when one considers that the 873,000 Liberal voters in the Prairies would have elected six more MPs. Meanwhile Conservative and NDP voters who were shut out of places such as the Atlantic provinces, metropolitan Montreal, Toronto, Winnipeg or the city of Vancouver would now have representation in all of those areas.

To view our full brief submitted the Special Committee on Electoral reform, please visit:

 1 With a 35% top-up, the size of local ridings would need to be increased by 54%. With a 40% top-up, it would be 66%. The formula is (1 – 1/1-Y) where Y is the share of top-up seats.


Return to all posts