Local district clusters vs Multi-member districts


There is a (possibly) new idea which had a fair bit of support in the preliminaries of the BC Referendum. It did not make it into the final 3 but a part of it is interesting and might be worth incorporating into other systems. This is the concept of local district clusters instead of Multi-member districts. The traditional Multi-member districts just take the single member districts and combine them. Lets take 5 for example. All single member districts are combined into clusters of 5 with 5 seats up for grabs and most major parties running 5 candidates. What the difference is in this “Local” system is that each candidate would be elected to a specific single member district inside the cluster of 5. So the first winner would eliminate all other candidates running in the district they won. To the best of my knowledge this was the idea of Byron Becker for his Local PR system (http://localpr.ca/basics/overview/).

Byrons system essentially uses this as a modification to STV so it gets all the problem which come with an ordinal system. However, this could easily be applied to something like RRV or PAV.


  1. Each elected person is responsible for a smaller area and can focus on the specific issues of that region
  2. Each citizen from that district has a specific representative for local issues who can be held to account. ie issues with traffic/schools/hospitals can’t have the buck passed by blaming other representative. For partisan issues they would go to the representative in their cluster who is most likely to agree with them like in STV.


  1. All the best candidates may be running in one district so when it is won they don’t compete in other races. The parties would have to decide who to run in what riding but I think this gets done already. I would favour a system where representative have to at least have a mailing address in the riding they hope to represent. In the last Canadian federal election an MP won a district without having a mailing address in Canada. So I am not sure this is really a con per se.

Different reweighting for RRV and the concept of Vote Unitarity

I know you’re saying it’s multi member, but what part of this is so? All I can see is multiple independent single winner districts. A side by side example would help.


I wonder if it’s possible to make a hybrid version of this and MMP, where you get the benefit of not doing party list offered by local PR, but without a district’s choice of representative appearing to be overturned by outsiders. A rough plan for this might look something like “elect the single range winner of each of the 5 districts, then elect 3 more with RRV, treating the 5 district winners as if they were elected in previous rounds.” (Or to incorporate your scheme, treating them as though they were elected with just the ballots in their district, and cap them accordingly.)
This probably has issues with strategy, though.


Hi @AssetVotingAdvocacy The thing you are missing is that you vote for all candidates in your cluster of districts. This means that in the end your vote may go towards electing a candidate in an adjacent district not the one you reside in. It is important to keep in mind that this happens in all PR systems, this is actually sort of what PR is.

@Marylander You could of course do this as a quick adaption but it would only obscure the effect. Lets say currently there are clusters of 5 districts. If we want to elect the 5 people for the local district directly with a single member system and then elect the next 5 with a multi-member system. This would mean that we would actually have to cluster 10 districts to get the 10 seats needed. But it would have to look like 5 so the districts would all double in size. This is something more like Dual Member Proportional but with 10 member instead of 2 member districts. I would rather just keep the basic method than do something like this because it is only about appearance. Imagine a version of STV where each elected member was randomly assigned to a sub-district for local issues. This would get us half way there.


So if I’m understanding correctly, voters will have influence over candidates outside of their cluster, but your hope is that most of the influence will come from voters within the cluster? This all seems unnecessary compared to letting candidates naturally declare “I want to represent XYZ area”, and letting voters decide if they want that or not. In other words, allow the voters to decide organically whether to pick a candidate based on common ideology, geography, or both! It isn’t as complicated as all this, and I can’t think of any examples where there’s a significant drop in quality over this.


You only have influence over districts inside your cluster. You influence districts other than your own but only in your cluster of districts. Candidates declare which district they are running in and only one wins per district. I think it is designed as you suggest.


It seems pretty good then. Mostly it seems like a political compromise rather than a real change to the system, since candidates will, as you said, shift around to be in the right districts. So if this modification makes PR look better, I’m all for it.


It is not all “political” there are real world benefits. I worked a lot with both the Pro PR and the No PR side in the BC referendum. One of the main issues people had with PR was a loss of accountability for the representatives. The two Pros I list in the original post are considered big wins by many people including myself. Other than the one Con listed the only other issue is the optics of influencing other districts. In most PR systems this effect is hidden.


But any voter who wants a geographic representative can get one. That’s the ideal of PR, and that’s how voters can get the accountability they seek under a regular multimember district PR system. There’s more “accountability” in this proposal, but at the cost of voter choice and perfect representativeness.


You must still be missing something because this would have no effect on the level of proportional representation or the voter choice.