MMP
Local ridings + proportional top-up seats
Understand / Proportional representation
How votes become representation.
Proportional representation, or PR, aims to make the share of seats a party wins reflect its share of the vote.
It is a principle, with different ways to put it into practice.
Canada · 2019
Votes and seats are different
Canada’s 2019 federal election: share of valid votes compared with share of the 338 seats.
Under first-past-the-post, each riding elects one MP. National seat totals add up those local victories. In 2019, the Conservatives received more votes nationally, while the Liberals won more seats and formed a minority government.
In Canada’s 2019 federal election, Liberal candidates received 33.1% of valid votes and won 157 of 338 seats, or 46.4%. Conservative candidates received 34.3% and won 121 seats, or 35.8%. The accompanying table provides the same comparison for every displayed group.
Liberal Party of Canada
Share of valid votes33.1%Share of seats46.4%Conservative Party of Canada
Share of valid votes34.3%Share of seats35.8%Bloc Québécois
Share of valid votes7.6%Share of seats9.5%New Democratic Party
Share of valid votes16.0%Share of seats7.1%Green Party of Canada
Share of valid votes6.5%Share of seats0.9%People's Party of Canada
Share of valid votes1.6%Share of seats0.0%All other candidates
Share of valid votes0.8%Share of seats0.3%
Show data table
| Party / group | VotesShare of valid votes | Seats wonShare of seats |
|---|---|---|
| Liberal Party of Canada | 6,018,72833.1% | 15746.4% |
| Conservative Party of Canada | 6,239,22734.3% | 12135.8% |
| Bloc Québécois | 1,387,0307.6% | 329.5% |
| New Democratic Party | 2,903,72216.0% | 247.1% |
| Green Party of Canada | 1,189,6076.5% | 30.9% |
| People's Party of Canada | 294,0921.6% | 00.0% |
| All other candidates | 138,4740.8% | 10.3% |
| Total | 18,170,880 | 338 |
“All other candidates” is a calculated aggregate of the remaining affiliations, independents, and unaffiliated candidates. It is not a party.
These are historical results, not a prediction of an election under proportional representation. Different rules could change how people vote and how parties campaign. Percentages are rounded; “all other candidates” combines several affiliations.
Sources: Elections Canada, Official Voting Results, 43rd General Election: Table 8, valid votes; Table 7, seats. Percentages derive from those integer totals.
01 / Explore the family
PR is a family of voting systems.
There is more than one way to achieve proportional representation. MMP and STV use different approaches to connect votes with seats.
STV
Multi-member ridings + ranked transferable votes
Mixed-Member Proportional representation
Let’s look first at MMP. It combines local constituency representatives with proportional top-up seats. The top-up allocation accounts for constituency seats already won.
MMP is a worked example of one proportional system here, not a statement of Fair Vote Canada’s preferred model.
Illustrative example
MMP: local seats plus proportional top-ups
A hypothetical 20-seat legislature shows how top-up seats account for local winners.
In this example, the party vote sets each party’s total share of the legislature. Constituency winners count toward that total. Compensatory seats fill the remaining places.
20 seats throughout: 12 constituency seats + 8 places reserved for compensation.
Step 1 of 3: Party votes indicate total seats.
L Constituency seats already wonT Compensatory seats
Party A
Party vote: 40%
Total seats indicated by party vote: 8
8 places indicated; 1 reserved for compensation.
7 local + 1 top-up = 8 total
Party B
Party vote: 35%
Total seats indicated by party vote: 7
7 places indicated; 3 reserved for compensation.
4 local + 3 top-up = 7 total
Party C
Party vote: 25%
Total seats indicated by party vote: 5
5 places indicated; 4 reserved for compensation.
1 local + 4 top-up = 5 total
Party A’s 40% share corresponds to eight seats. Seven constituency winners already count toward those eight, so Party A receives one compensatory seat.
Read the complete example
Party A receives 40% of party votes: seven constituency seats plus one compensatory seat, eight total. Party B receives 35%: four plus three, seven total. Party C receives 25%: one plus four, five total. The legislature contains 20 seats.
This simplified example has no threshold, rounding problem, or overhang. Real MMP systems differ in their lists, regions, seat numbers, and treatment of constituency results.
Illustrative example based on the compensatory principle of MMP. See Elections New Zealand and ACE Electoral Knowledge Network.
02 / Look at the design
Local representation and design choices.
Local and regional connections
MMP retains constituency representatives while improving party proportionality. Regional lists can connect top-up representatives to places. Smaller regions can reduce proportionality.
Choices and trade-offs
Keeping the same legislature size generally means fewer, larger constituencies to make room for list seats. Thresholds and inadequate compensation can still exclude or underrepresent voters.
Background reading: ACE’s MMP explanation ↗
03 / Ask questions
The details matter.
Would I still have local representation?
A PR design can include representatives elected from local or regional districts. The geographic scale and number of representatives depend on the model chosen.
Is a ranked ballot always proportional?
No. Ranking describes how voters express preferences. Whether a result is proportional also depends on how many seats are elected together and how those seats are allocated.
Would votes and seats match exactly?
Not necessarily. Seat counts are whole numbers, and district sizes, thresholds, and counting rules can all affect proportionality. The exact match in our visual is deliberately simplified.
Does this prototype propose a specific system?
No. It provides an introduction and a structure for exploring evidence. An expanded version could compare specific models and their tradeoffs in more depth.