Local Transfer Vote(LTV): A weird way to attempt PR

On my journey with researching Proportional Representation I’ve always leaned towards the ranked methods of Single Transferable Vote, and it’s variants also known as STV+ (STV + MMP “Top up seats” and Rural Urban Proportional). Recently I’ve come across variations of the rules trying to make rules similar or more fair such as having a province wide top up seat that lines up with 1 seat per riding, or having the ranked choice on parties with preferred member to be selected for parliament.

It reminded me of one suggested system that was to combine Alternative Vote (AV)(Single Member ranked choice riding), and STV.

Local Transferable Vote was one of the submissions for the Electoral Reform Commission of 2016. LTV is a variation of STV that has a dual layer of AV/STV voting. Single Member districts would remain for local using instant run off, and using STV as a top up. If a candidate does not get 51%  locally they have a chance of being regionally selected depending on a region’s quota. The number of local seats to region attempts to have a 6:4 ratio of seats per region similar to most MMP compositions. So a region of 10 seats would have 6 local seats and 4 regional seats. The regional quota uses droop which is the remaining percentage of the non-local vote is split among the plus 1 plus 1 vote.

page4image2216176576

As a result you have a ballot which labels your local candidates the regional candidates in the other regions. You would rank as many candidates as you had an interest in. When the voting happens each on the local level a candidate must get over 50% of the vote to be elected, regionally a candidate must get above the quota depended on the regional composition we mentioned before. Candidates are either elected or eliminated until there are enough seats both locally and regionally.

To get elected, a candidate must meet at least one of the following conditions: 1) Win more than 50% of the first choice votes in his local constituency.

(First choice votes from other constituencies wouldn't count towards this goal.) 2) Meet the regional quota with either first-choice or transferred votes, region-wide. 3) Become the last candidate from his local constituency to remain in the count.

The counting would proceed as in any other preferential voting system. Unless protected by condition 3, the candidate with the least votes would be eliminated and his votes transferred in accordance with voters' second or subsequent choices. This process is repeated until all the seats in the region have been filled.

In the event of a surplus (when the candidate's vote count exceeds the regional quota), the last batch (that has caused the surplus) would be recounted and each vote would be transferred at a value equal to the number of surplus votes divided by the number of ballot papers received in the last transfer.

To ensure that all seats are won with the same number of votes (or as close to that as possible), candidates would remain in the count until they actually meet the regional quota, even if they have already been elected under conditions 1 or 3.

Since the regional vote is STV this means the vote can be transferred to independents regionally as well as locally. Something that most MMP Systems would not allow (unless there is a sort of alliance list). One of the most interesting things the paper recommend is for Unaffiliated candidates to self designate their positions to help distinguish themselves. This is pretty unique though I can total see this used for memery.  And that makes me smile.

page11image2218156768

And that’s LTV. It’s an interesting idea for electoral reform in Canada, and implementation of an STV system. I can see this being a be appeal to independent and STV fans. I do have to wonder how an electoral map would look if this was implemented. There are examples describing how these districts might be draw in the paper, but the question does rise of how local is a local constituency. Also like Rup a design decision is to main the amount of seats in parliament but have a portion be for the regional mps. Maintaining the 6:4 ratio the parliament make up would 210 local : 128 Top up. With 35-40 Electoral top regions the author claims that this would be a sweet spot between MMP proposals, and P3 or STV with regards to regions and proportionality.

Though still way more than the 50 top up that RUP and other STV+ Proposals.

Additional note: Surplus Transfer differences:

An aspect of any stv system is how to transfer surpluses. There’s a few methods but the most commonly proposed is Gregory method (WIGM) which is Surplus votes/Total Candidate votes. What this means in practice is each surplus transfer is fractionally transferred the further you go in the rounds of counting.

However the author of LTV appears to recommend a Hare-clark method which is Surplus votes/the last batch of votes to cause that surplus. So if a second preference caused a surplus only the second round of votes would be in that transfer ratio. Or the other way, each preference is it own count per surplus.

These obviously will cause different results. The Method of how a surplus is an important discussion when talking about implementing a STV like system and the discussion of which is best.


A Nemes Content Blog 2022.