This year, BC is ground zero for electoral reform in Cascadia. A fall referendum presents provincial voters with a major opportunity to update their system and make their voices heard and to introduce proportional representation (ProRep) to the big leagues on this continent. 

Any of these options would be better than first past the post because they would make more votes matter, achieve fairer electoral results, and generally bring BC into the 21st century of the democratic world.
Tweet This

The choice BC voters will see on their mail-in ballot this fall is something like this: Imagine you are faced with options behind Door A and Door B, but the doors are cracked open so you can see what’s beyond. Behind Door A is a shack without indoor plumbing or lights. Behind Door B are three options, each of them a well-constructed, modern house complete with 21st century amenities like running water and electricity.  

As I explained last time, voters will first get to choose between Doors A and B. Behind Door A is First Past the Post (FPTP), an outdated 18th century voting system in which the candidates with the most votes wins, even if that candidate won less than half the votes. Also known as winner-take-all plurality voting, first past the post leaves large swaths of voters with no voice in government (because they didn’t want the one winner who represents their whole area) and often leads to the wrong party gaining power, as in five of the last seven BC elections when the party with fewer votes still took total control in Victoria. 

Behind Door B is proportional representation (ProRep) a system that most developed countries in the world use to make more votes count and make sure that only majority support from voters can lead to majority control of government. Proportional means that the number of seats reflects the share of votes. If voters choose Door B, they will then get to decide among three versions of ProRep. Each has pros and cons, but every one of them would be a huge upgrade from the shack of first past the post.  

Today, I will lay out what’s known about the three proportional options BC voters will consider. Any of these options would be better than first past the post because they would make more votes matter, achieve fairer electoral results, and generally bring BC into the 21st century of the democratic world. Although there’s a lot to say about each system, the detailed design of the winning option will happen after the referendum is done. A committee composed of representatives of all the parties in the BC legislative assembly will turn the outline the voters choose into a law. At least, that’s the process that BC Attorney General David Eby recommended last month in his official report to the province. Until the BC Cabinet approves (or modifies) Eby’s recommendations, even that much remains only tentative. 

Still, Eby’s recommendations are thorough and important, so I will spell out what they mean. First, an overview of the three ProRep systems. Two are made in Canada. All three are better than First Past the Post. 

  • Mixed Member Proportional (MMP) is a well-proven system that retains many of the features of the current system. 
  • Dual Member Proportional (DMP) is a “made in Canada” variant of MMP with a simple ballot where voters fill in just one bubble per race, like now.  
  • Rural-Urban Proportional Representation (R-U PR) is a “made in Canada” system that gives voters the chance to pick their candidate, without letting parties pick for them. It uses ranked ballots—a well-proven system recommended by the 2004 Citizens’ Assembly—in urban and suburban areas, and MMP (see above) in rural areas.  

Better than First Past the Post 

Many Canadian and American voters have started to realize that old FPTP voting leaves a lot to be desired when it comes to fair representative democracy. Often, the winning party gets less than half the votes, but all the power. A party with less than half the votes won control of the government in 10 of the past 12 provincial elections in BC. Individual votes don’t translate into representation, so voters feel frustrated and Members of the Legislative Assembly (MLAs) aren’t always accountable to voters.  

This fall, BC voters will have the chance to choose a better system. BC Attorney General Eby recommends that, whichever system the voters choose, it be designed to meet three criteria voters told him were important during an extensive process of gathering public feedback:  

  1. Keep the same number of Members of the Legislative Assembly (MLAs)—87, or increase to a maximum of 95 members, 
  2. don’t dilute regional representation, and
  3. don’t allow fringe parties with less than five percent of the vote to win seats.  

The sections below describe the three systems, where they are used, and an example for each of one way the system might work in practice in BC. (Just an example, since the legislative committee will design the final details). The sections below also describe how the three systems vary on several additional properties that have arisen during past discussions about electoral reform in BC, and during the recent public comment period that Eby’s recommendations took into consideration: 

  • Votes count. BC voters want to know their individual vote matters. One way to measure this is by the percentage of individual voters who cast a vote for a specific MLA who now represents them.  
  • Fairness. If ten percent of voters resonate most with the Green Party, it is fair for the Green Party to win ten percent of the seats.  
  • Ballot. Voters want the ballot to be easy to understand. 
  • Geographically compact rural districts. Some rural ridings (or “districts,” in American parlance) are geographically huge. For example, Stikine, BC’s geographically largest riding, is 196,484 square kilometers (or 75,863 square miles, larger than 33 American states). It takes the MLA 21 hours to drive across his jurisdiction. For MLAs to do their job, a few of these rural ridings really can’t get much bigger.   
  • Voters choose vs. parties pick. BC voters want to pick their MLAs themselves, not have a political party choose for them.  

Voters who want a well-proven system that retains many of the features of the current system may like MMP the best. 

Germany, New Zealand, Scotland, Wales, Bavaria, and several other countries use MMP to elect their national legislatures. (Videos here and here explain the forms of MMP used in New Zealand and Germany.) In each case, the countries see higher rates of representation for political minorities, women, and people of color. The eastern Canadian province of Prince Edward Island voted in 2016 to adopt Mixed Member Proportional Voting to elect its provincial legislature. Unfortunately, the measure was non-binding, and the legislature ignored it. 


In MMP, people get two votes. The first is almost the same as in past BC FPTP elections, just with districts that are a little bit larger. That is, you vote for one candidate in your local riding, and the candidate with the most votes wins, whether that candidate got 60 percent of the vote or 40 percent.  

The second vote is where things get interesting. With your second vote, you’re invited to pick a regional representative (see the example ballot below) or a favorite party (the legislative committee will decide this detail). If you vote for an individual, your vote will count in two ways, both for the individual candidate and for that candidate’s party. I’ll explain how this works in a moment. If you vote for a party, that party will assign their regional seats based on a party list. Note that voters in New Zealand and Germany vote for a party with that second vote, and the parties use country-wide lists. Bavaria runs regional candidates, which is what Eby recommends for BC.  

At least 60 percent of the MLAs (the legislative committee will decide the exact percentage) would be elected as they are now—using First Past the Post (FPTP) in local ridings. Because the number of locally elected representatives would be smaller than the current 87, each district would have to be somewhat larger.  

The remaining seats—up to 40 percent of them—would be reserved for regional representatives. The regional representatives would round out the regional delegation to ensure that the legislative assembly, overall, would fairly reflects voters’ party preferences as expressed with their second vote. First, all the regional votes for each party would be tallied up. If voters get to cast votes for individual candidates, those would all count towards the party—a vote for Jane Smith of the NDP would count as one vote for the NDP and so would a vote for John Jones of the NDP. The votes for candidates of each party would be divided by the total votes cast in the province to come up with each party’s percentage. Each party’s percentage of total seats (local and regional) should equal its percentage of total votes. Each party’s local winners would be subtracted from their share of seats to determine how many of the regional seats each party wins.  

The example in the box walks through some hypothetical numbers.  

Original Sightline Institute graphic, available under our free use policy.


Say BC were to increase its legislative assembly to 95 (the maximum Eby’s recommendations allow). It could draw 60 single-winner ridings. Residents of each district would elect a representative for their riding. These 60 ridings would also be grouped into five equal-population regions each made up of twelve ridings; each region would elect seven regional representatives. Voters would vote for their favorite local district representative and their favorite regional representatives. The region as a whole would send 19 representatives to Victoria—twelve local MLAs and seven regional MLAs—and any given voter would have 8 MLAs representing her in parliament.  

For example, say the party shares of provincial votes were 42 percent NDP, 38 percent Liberal, 14 percent Green, and 6 percent Conservative. And say NDP and Liberals each already won 6 of the 12 local ridings in a region and Greens and Conservatives won none, so NDP would get 2 of the regional seats to bring its total up to 8 of the 19 regional seats (42 percent); Liberals would get one regional seat to bring their total up to 7 of 19 (37 percent); Greens would get 3 regional seats and Conservatives 1. Candidates would fill the regional seats in order of their individual popularity—the two regional NDP candidates with the most individual votes to their names would fill the NDP’s two regional seats. 

As a voter, say you voted for an NDP candidate for your local riding representative, and your favorite Green Party candidate from the regional list. The Liberal candidate won your local riding. Under FPTP you would have been out of luck with no MLA you voted for heading to Victoria. But under MMP, your second vote still counted, and your favorite Green candidate got one of the regional seats and is now working hard for you in Victoria.  

  • Votes count: About 30 to 70 percent of voters would have an MLA they specifically voted for. About 30 to 50 percent of voters would vote for a winner in the local ridings (because the candidate with the most votes would win, but with more than two candidates the most votes could be one-third or less). Somewhere around one-third of the remaining voters would vote for a candidate who would win a regional seat. The rest of voters might still have their regional vote count towards that party winning a regional seat, though not their preferred candidate within the party.   
  • Fairness: MMP guarantees very proportional outcomes, so voters’ party preferences are closely reflected in Victoria.
  • Ballot: Half the ballot is the same as now, and the other half is still a simple “vote for one” format, but that vote counts for both the individual candidate and their party. 
  • Geographically compact rural districts: All voters would still have a local representative, though the local ridings would be larger than they are now 
  • Voters choose (vs. parties pick): If voters are able to vote for an individual candidate for the regional seats, they would be able to choose a favorite candidate and that vote would help that candidate win a regional seat. But if they are only able to vote for a party, the party would get to dictate the list of candidates to fill the regional seats. 

For voters who want a simple ballot, DMP is an attractive choice with its basic “vote for one” structure. 

Dual Member PR (DMP), also known as Dual-Member Mixed Proportional, is a “made-in-Canada” variant of MMP. As discussed above, many countries have used MMP for decades and more than a dozen countries use a system of PR called Closed Party Lists, which is similar to DMP. So, while DMP itself has never been used, it is based on two well-tested systems. Some Canadians are very interested in the system. For example, it appeared on Prince Edward Island’s 2016 ballot about electoral methods where it garnered the third most votes of the five options PEI voters had to choose from. The group DMP for Canada has used actual votes from provincial and federal elections in Canada to simulate how the election would have turned out with DMP. 


With DMP, voters would see a simple “vote for one” ballot (example below), but each line would list one party and up to two candidates for that party, and your vote for that line would count both for the candidates and for their party. Most ridings would be about twice the size of BC’s current ridings and would elect two MLAs, except that a few rural ridings might remain about the same size as they are now and elect one MLA, as they do now.  

The first MLA elected in each riding would be the first candidate listed for the party that won the most votes in the riding. For example, if the Liberal ticket in a particular riding won 40 percent of the votes, and that was more than any other party’s ticket, the first-listed Liberal candidate would win a seat in the riding.  

Determining the second winner in the riding depends on the votes in the riding, and on the votes province-wide. The second candidate on the winning party ticket gets half the votes for that party. In our example above, where the Liberals got 40 percent of the vote in the riding, the first Liberal gets a seat and the second Liberal gets 20 percent of the vote. Elections BC would then make a list of each party’s remaining candidates (the first-listed in riding where that party did not win, and the second-listed in ridings where they already won a seat) in order of the percentage of votes they won in their riding.  

  • Our work is made possible by the generosity of people like you!

    Thanks to Sherry Richardson & James Montague for supporting a sustainable Northwest.

  • Elections BC would calculate the percentage share of province-wide votes that each party received, and the corresponding number of seats each party should win, and subtract the number of seats each party already won to get the number of seats that party still needs. Elections BC would then select that number of candidates from the top of the party’s ordered list. (A video explains using the example of Alberta, here).  


    Say BC continued with its current 87 MLAs, keeping the three geographically largest (most sparsely populated) rural ridings the same as now, and combining the remaining 84 ridings into 42 double-sized local ridings. Each local riding would elect one local representative in a single-winner first past the post race. The rural ridings would work just like now, and the first winner in the double-sized ridings would be the first candidate listed on the line that wins the most votes in that riding, for a total of 45 MLAs.  

    The remaining 42 seats would be distributed to a second candidate in each double-sized riding based on how many seats each party needs to proportionally reflect the province-wide vote for that party, and how well the parties did in the local race.  

    Say Liberals won 41 percent of the province-wide vote, NDP 40 percent, the Green Party 16 percent, and Conservatives the remaining 3 percent. Conservatives didn’t hit the mandatory 5 percent threshold, so they would not win any seats. Instead, Liberals would get 37 seats, NDP 36, and Greens 14. Say Liberals already won 19 of the local ridings, so they get 18 of the remaining 42 seats. Elections BC would list each parties’ remaining candidates (their second-listed candidates in ridings where they already won a seat, with the second-listed candidate getting half the votes, and the first-listed in ridings where the party did not win) in order of percentage of votes the party ticket won in that riding. The top 18 Liberal candidates—those that won the highest percentage of votes in their riding—would win seats. Elections BC would repeat this process until each double-sized riding had two MLAs and each party had their fair share of seats. 

    Original Sightline Institute graphic, available under our free use policy.

    • Votes count: Assuming voters are directly voting for the first candidates listed on the ballot, somewhere around 40 to 70 percent of voters would have an MLA they specifically voted for. About 30 to 50 percent of voters would vote for a winner in the local ridings (because the candidate with the most votes would win, but with more than two candidates the most votes could be one-third or less). For the remaining half of seats, probably most, say around three-quarters, would go to a party that had not won the first seat in that riding, and by definition less than half the voters in that riding would have voted for that party’s pair. The remaining seats would go to the second candidate in a riding where that party was extremely popular, but no voters directly voted for the second listed candidates.   
    • Fairness: DMP guarantees very proportional outcomes, so voters’ preferences are more accurately reflected in Victoria 
    • Ballot: the ballot is a simple “vote for one” ballot like now, but that one vote will count for two candidates and their party 
    • Geographically compact rural districts: The largest rural districts would remain unchanged, so they would not get larger. All other districts would be twice as big as now. 
    • Voter choice within or across party: Voters would have to accept the candidates in the order the party chose. If a voter preferred the second-listed candidate, he would not be able to indicate that. 

    For voters who want more control, R-U PR is the best choice because it lets voters vote for or rank candidates directly, with very little intervention from political parties. 

    Rural-Urban Proportional Representation (R-U PR) is a “made-in-Canada” solution aimed at fairly representing both rural and urban populations without swelling rural districts to unmanageable sizes. It uses Single Transferable Vote (STV) in urban and suburban areas. STV is used in Cambridge, Massachusetts, Ireland, Australia, local elections in Scotland and New Zealand, and for Academy Awards nominees. In 2004, the British Columbia Citizens’ Assembly on Electoral Reform recommended the system for BC elections and in 2005, 58 percent of BC voters voted to adopt STV. Unfortunately, that referendum needed 60 percent of the vote to go into effect. 

    Elections in rural areas would be slightly different from those in more populated areas—they’d use  Mixed Member Proportional (MMP) voting to ensure local representation in these areas. As discussed above, several countries including New Zealand have long used MMP.  


    Voters in suburban and urban areas would see a multi-winner ranked ballot (see an example below). They would elect two to six representatives per riding, depending on how densely populated the riding was.   

    Voters in rural areas would see an MMP ballot. They’d “vote for one” to elect a local MLA from a local riding about the same size as now. And they’d also cast a second vote for a favorite rural regional candidate (or maybe for a regional party, depending on what the committee decides) that would count for both that individual candidates and that candidate’s party. The rural regional seats—between 7 and 13 seats total—would match the party preferences of the rural voters, and also honor their individual candidate preferences. As described in the MMP section above, Elections BC would add up all rural regional votes to see what percentage each party won, then subtract the number of local ridings each party already won, and allocate the regional seats to the most popular rural regional candidates from those parties.   

    Original Sightline Institute graphic, available under our free use policy.


    For example, say BC increased its legislative assembly to 95 (the maximum Eby’s recommendations allow). It could elect: 

    • 60 MLA’s from urban areas—10 urban ridings with 6 MLAs each 
    • 15 MLAs from suburban areas—5 suburban ridings with 3 MLAs each 
    • 12 MLAs from rural areas—13 rural ridings with 1 MLA each 
    • 8 regional rural MLAs, to ensure overall proportionality 

    Urban and suburban voters would get to rank candidates in their own order of preference, regardless of what order the parties would prefer them in. The three most popular candidates in each suburban riding would go to Victoria as would the six most popular in each urban riding. For these regions, there would be no party lists and no calculating of party’s share. But the overall results would very closely reflect the will of the voters, because the voters had the chance to express their exact order of preference for individual candidates.  

    Rural voters would elect a local MLA just like now. They would also get to cast a vote for a regional candidate. Say 50 percent of rural voters chose a Liberal regional candidate, 30 percent chose a conservative, and 20 percent chose an NDP. If conservative candidates won at least 5 percent of the vote province-wide, then they would get 6 of the 20 rural seats, the Liberals would get 10 and the NDP would get 4. If Liberals already won 8 of the rural local ridings and Conservatives already won 4, then the 2 most popular Liberal regional candidates and the 6 most popular conservative rural candidates would win the rural regional seats.  

    • Votes count: About 75 percent of BC voters would have an MLA they specifically voted for. In 6-winner ridings, at least 85 percent of voters would have an MLA they ranked; in 3-winner ridings, at least 75 percent would; and in rural ridings around 65 percent would. 
    • Fairness: R-U PR guarantees the fairest outcomes. Voters’ preferences would be very accurately reflected in Victoria. 
    • Ballot: for rural voters, the ballot would be a familiar “vote for one” ballot like now. For urban and suburban voters, it would be a ranked ballot. 
    • Geographically compact rural districts: Rural districts would stay the same size as now. 
    • Voters choose (vs parties pick): R-U PR maximizes voter choice and minimizes the role of parties—voters in urban and suburban ridings could rank any candidates they wished, regardless of party. 

    What’s a BC voter to think about all these options? 

    BC voters, like most North Americans, are living in an electoral shack—an outdated 18th century model without any of the modern amenities. BC voters have seen the drawbacks of first-past-the-post first hand, with election after election going to the party that won fewer than half the votes, and sometimes fewer votes than another party. Votes go to waste when they don’t actually translate into seats, nor into control of government. 

    But BC voters could change all that, this fall, and light the way for other American and Canadian jurisdictions while they’re at it. Any of the three proportional systems on the BC ballot would be a step up from first past the post. They are all modern electoral systems that make more votes count and achieve fairer results.  

    For voters who want more control over their ballot, R-U PR is the best way to go because it lets voters vote for or rank candidates directly, without parties intervening. For voters who would prefer a simple ballot, DMP is the best choice with its simple “vote-for-one” structure. For voters who want a tried and tested model, MMP is there for you with a track record in New Zealand and elsewhere that spans decades. But the important thing is, voters will have the chance to decide whether to replace their outdated system with something better.