In choosing one or more proportional voting systems to appear opposite First Past the Post (FPTP) on the upcoming BC referendum on electoral reform, one must sooner or later consider concrete options. This document enumerates eight possible systems and provides a one-page description of each.
Download the full catalog of voting systems (PDF, 12 pages)
The eight selected systems are listed in the table below. Five are forms of Mixed Member Proportional (MMP) that would result in a two-tier legislature consisting of district MLAs and regional MLAs (Systems 1, 2, 3, 5, and 7); the other three maintain a single tier of local MLAs but associate them with multi-member districts. Six of the models use a single-vote ballot (Systems 1 to 6), one introduces a 2nd vote (System 7), and one employs a ranked ballot (System 8). Four of the options entail competition among candidates of the same party (Systems 5 to 8).
| System | Name | Number of Regions | Tiers | Ballot | Intra-Party Competition |
| 1 | Closed-List MMP with FPTP ballot | 4 to 8 | 2 | 1-vote | no |
| 2 | List-Free MMP with FPTP ballot | 4 to 8 | 2 | 1-vote | no |
| 3 | List-Free MMP with secondary candidates | 4 to 8 | 2 | 1-vote | no |
| 4 | Dual Member Proportional | 1 to 5 | 1 | 1-vote | no |
| 5 | List-Free MMP with rankable candidates | 4 to 8 | 2 | 1-vote | yes |
| 6 | Dual Member Proportional with rankable candidates | 1 to 5 | 1 | 1-vote | yes |
| 7 | Open-List MMP with Bavarian metrics | 7 to 8 | 2 | 2-vote | yes |
| 8 | Single Transferable Vote | 16 to 22 | 1 | ranked | yes |
By submitting eight options for consideration, one of my objectives is to steer decision-makers away from inadvertently introducing ticket-splitting tactics into provincial elections. These tactics emerge under certain two-vote MMP models which exclude district votes from the calculations that allocate compensatory seats to parties. Under such systems, many British Columbians would have to endure appeals to give their district votes to a large party and their regional votes to a small party. These appeals would be based on the expectation that the small party could not win at the district level, whereas the large party may win so many district seats that it receives no regional seats. The eight selected systems minimize or avoid tactical considerations of this nature while collectively offering a diverse set of possibilities for voting and representation.
The system descriptions are followed by a discussion, a glossary, and recommended reading.
The catalog was prepared with contributions and editorial assistance from Charles Cares. Click the link near the top of this post to download the PDF.
Rhys, this is excellent. But I am curious.. In constructing your catalog, you have ruled out list PR with moderate district size (EG M=~4 as in your STV case). What is your rationale for this?
(I too would exclude or deprecate list PR. In my case it would be because of its awkward tension between proportionality and district magnitude, with both of them ending up having to be fudged. But I would also exclude or deprecate STV, for the same reason.)
At a more mundane level, I don’t understand the ‘Regions’ column in your summary table.
Hi Chris,
In my opinion, List PR only works well if you can ensure a proportional outcome. This means you need about twice as many seats as competitive parties in each district. BC has 3 competitive parties, but there could be 4 with proportional representation. So you need at least 8 seats per district, which may not go over well in BC’s referendum.
List PR with only 4 seats per district might perpetuate tactical voting. For example, suppose your favourite party is polling at 7% in your district. They won’t win any seats, so voting for them will have no effect. Suppose your 2nd favourite party is polling at 27%. Chances are they will win exactly 1 seat, no more and no fewer, so voting for them will again have little effect. Suppose your 3rd favourite party is polling at 17%. They may or may not get a seat, so this is the party you should vote for if you want to maximize your influence. But this is your 3rd choice!
Parties will also realize that additional votes are worth less in certain districts, so they will be less accountable to potential supporters in those districts.
Actually STV should work much better than List PR in 4-seat districts. You generally don’t need to think about voting tactically, as your vote will automatically transfer to where it is most beneficial. Also, one or more seats in every district will be hard to predict, so most parties have at least some motivation to pursue support across the province.
Of course there are other advantages and disadvantages of STV. For me, the bottom line is that it was the consensus model of the randomly selected members of the BC Citizen’s Assembly.
I just changed the column heading from “Regions” to “Number of Regions”. The entry “4 to 8”, as an example, means that BC would be divided into 4-8 regions. Election results would only be proportional within each region, not province-wide. In List PR and STV, regions and districts are the same thing. For MMP and DMP, there are multiple districts within each region.
Hi Rhys, I don’t have references at hand but studies indicate most voters don’t vote tactically. For example, when they could maximize their influence by voting for a large party on a local ballot, and a small party which is their true first preference on a regional MMP ballot, the great majority simply vote for their favourite party across the board, instead of voting tactically. So among the very many factors that need to be well balanced in a good voting system, avoiding tactical maneuvering that might give some voters slightly more potential influence than others would seem to rank pretty low. Rather than choose the eight candidate systems on that basis, it might be better to choose eight on the basis of the voters’ perceived top priority, which is simplicity.
Also, any voting system with reasonable proportionality should minimize the common vote splitting problem of FPTP.
Hi Maxwell, I understand and appreciate your points.
Perhaps. It’s important to look specifically at Scotland and Wales, which use MMP models similar to what will probably be put forward in BC. In Scotland, small parties often do not run at the local level, so it is hard to tell whether someone who splits their ticket honestly prefers the small regional party or supports them out of perceived necessity. In Wales, a glance at recent voting patterns suggests to me that there is a fair amount of tactical voting.
But either way, just having an incentive to vote tactically can be a serious problem whether or not it changes how a person votes.
In First Past the Post, a Green supporter may vote Green even if a different party has a better chance of defeating her least favourite candidate. Her decision to vote honestly doesn’t undo the fact she faced a difficult and unnecessary dilemma. Similarly in MMP, one might give their list vote to a regionally dominant Conservative party even though a smaller like-minded party has a better chance of winning a compensatory seat. Again, voting one’s conscience doesn’t solve the underlying problem. Tactical incentives are associated with ineffective votes and reduced faith in the electoral system. It would be great to get rid of voting strategies specific to FPTP, but why introduce new voting strategies in the process?
Many would agree with you, but I see no downside to avoiding MMP models that are vulnerable to ticket splitting tactics. If you want a simple ballot, you can have it with a tactic-resistant single-vote MMP model (see Systems 1-3). Or if you want two votes, you can have that too while avoiding risk (System 7).
That’s a fair argument. Of course it’s hard to get those who know voting systems to agree on the best type of ballot, the ideal number of MLAs per district, whether to introduce a 2nd tier of representatives, etc. I believe we should all agree to try to minimize the risk of introducing new tactical considerations into the voting system. My main point is that we can minimize this risk without severely limiting our options.