[EM] Preferential voting system where a candidate may win multiple seats

Juho Laatu juho.laatu at gmail.com
Mon Jul 8 03:01:35 PDT 2013


On 7.7.2013, at 23.49, Kristofer Munsterhjelm wrote:

> But this raises the question of where the regional MPs should reside.

Two approaches (just thinking out loud). 1) One could have multiple layers from single member districts to counties etc. I recommend natural historical borderlines, not newly generated random borderlines, since it is important that people see the regions and their representatives as their own, representing some identifiable group with similar needs and targets. 2) Each party could have their own districting that spreads their representatives over all districts (smallest parties with one representative would have one district that covers the whole country).

I note that in order to keep good geographical balance, one would have to take into account where each representative comes from, and not let their distribution become too unbalanced (e.g. one bottom level district gets representatives at all levels and another one only at the bottom level). This seems to get quite complex.

> And yes, when one adjusts local outcomes to get greater national proportionality, that means that someone who "shouldn't" have won on the local level nevertheless does win. Hopefully the difference won't be as great as to make the voters complain! Perhaps this is part of the point of leveling seats: they start off not being owned by anyone, so giving them out to party members may not seem as much a way of overruling the local result as if one started with all seats filled and *then* adjusted.

Yes, it is important to give also a good impression of how the system works.

I think people are quite ok with the idea that there will be some randomness in the allocation of the last seats, since that is the case anyway in most methods, and since those candidates are anyway close to having vs. not having sufficient support. Better luck next time for those candidates that didn't make it this time.

I note also that one easy trick to get accureate proportionalities (national political, regional political and geographic) is to have representatives with different weight (different voting weight while in the parliament). But usually people do not fancy this kind of solutions.

I'll outline also one sketch of a simple non-backtracking algorithm, just for reference. 1) Allocate (number of) seats to parties at country level, 2) in each district, allocate those seats that are supported by full quota of party votes, 3) allocate the remaining seats to parties, starting from the smallest party, so that each party gets its seats in regions that still have unallocated seats left and where the party has highest support, 4) allocate the seats of each party in each district to their candidates. Point 3 is the critical one. The idea is simply that small parties better be handled first since large parties have probably good candidates with reasonable amout of support in each district, and they can therefore be allocated last (without causing strange results). The result may not be ideal, but probably good enough, and acceptable since the algorithm is simple and straight forward.

Juho






More information about the Election-Methods mailing list