How to run a PR top up system with multi-seat constituencies

The point of multi seat constituencies is to provide the electorate with a choice not alone of party but of personnel. It is in effect a nascent instant primary system. However political parties don’t see it like that and tend towards running the minimal slate of candidates because they are currently fearful of using it as such because of the effect of transfers on the number of seats they


What if that problem was to be corrected by the allocation of the total number of seats in the Dail on the basis of your share of national 1st preferences and not on the basis of PR-STV but that PR-STV was instead used to decide who would elected.

Leaving aside the idea of reducing the size of the Dail itself as it complicates the matter I would suggest the illustrative below.

We would have 123 of the existing 166 TDs elected by from the multi-seat constituencies as at present but once the counts have completed that the remaining 43 of 166 seats would be allotted to the parties smallest first on the basis of which of their candidates had the highest number of votes before being eliminated or perhaps not eliminated but simply not elected (the last man standing scenario). It would be proportionate yet also ensure that the public were deciding who was getting elected and not some closed insider type party list system. And it wouldn’t hurt parties if they ran 8 candidates in a 5 seater which would mean more choice for the voters.

I might try and see how the last Dáil might have been changed with this system.

  • joeCanuck

    Looking forward to seeing your analysis, Dan. It sounds an interesting concept and I believe it is used for national elections somewhere – Israel?

  • Hard to run a simulation of that sort of system based on Dail elections since there are 3, 4 and 5 seaters. You’re not providing much detail on your methodology here. Are you suggesting to reduce all constituencies by one seat and then run a top-up mechanism like d’Hondt(?) based on overall first preferences?

  • Dan Sullivan

    The lazy approach (which I might well adopt for the sake of simplicity) would be to reduce each of the constituencies by one seat. And yes, I’m suggesting that we allocate the remaining 43 seats (that proportion could be altered) of the 166 on the basis of d’Hondt (at least I think it would be d’Hondt but sometimes people say stuff to undermine my confidence that I understand d’Hondt properly).

    Of course the point of this system would be to entice more people from a broader range of backgrounds to stand for election as the negative impact of potentially lost transfers wouldn’t be costing parties seats any more. I’ve been toying with the idea for a while and was prompted to write as it seems a better alternative to the notion of adopting gender quotas for elections.

  • aquifer

    The Additional Member System does something like this so that the total result ends up more proportional, and also gives national ‘communities of interest’ or minorities more of a chance to get elected.

  • Dan Sullivan

    It has similarities with the additional member system but it removes the control from the party hierarchy that a list from which the additional members would be drawn. In this system the additional members are also people who stood in the PR-STV vote.

    One of the recurring arguments against lists or top ups is that it places more power in the hands of parties at the expense of ordinary voters. And I would have concerns about that too, there is considerable favouritism and in part nepotism in politics in who gets chosen for tickets. This is an attempt to remove that as an argument against a top up system.

  • One of the arguments FOR an additional member system in the Republic is having some elected reps in the Dail who have no consituency base (and all that seems to come with that in Dail Eireann). The additional power this accords the parties may be a relief compared to having to select the correct mix of ministers/junior ministers/opposition front bench spokepersons etc. Of course, that correct mix is indexed against winning and losing seats etc and rarely does aptitude or competence feature as the defining criteria.
    Don’t envy you trying to simulate the results based on the 2007 elections. Just looked at the first five:
    Carlow-Kilkenny would see FG take the Green seat by either electing the last seat by d’Hondt (based on party prefs) or the highest unelected First Pref vote.
    Cavan-Monaghan had the Ceann Comhairle and would need special rules to simulate! It would probably have elected 2 FF, 2 FG and 1 SF based on overall party results (i.e. 1 FF loss to FG).
    Clare would produce same result whatever way.
    Cork East would produce the same (via d’Hondt) or Lab loss/FG gain if the fourth seat went to the unelected candidate with the highest First Pref.
    Cork North Central would produce same result whatever way it was run.
    At first glance, smaller parties may lose out overall to the likes of FF and FG (Labour are traditionally believed to enjoy a considerable bounce in seat numbers relative to their support). Independents, in particular, would be the most likely to miss out. Overall, the system would discourage the arms length FF/FG gene pool TDs to maximise party strength for the last seat (with whatever deals needed to keep them onside). Basically this system effectively encourages conformity.

  • John, I agree with then need for national politicians who aren’t tied to the constituency. This is more a working out of a system rather than my personal endorsement of it as the solution to all our problems.

    As for the example you use, I think you’re applying d’Hondt individually to every constituency one by one

    My idea would have given FF get 69 seats, FG 45, Labour 17, SF 11, Greens 8, PD 4 SP 1 and PBP 1 and independents getting the remaining 10 seats (I have very mixed feelings about the allocation to independents) as that is what their 1st preference share was.

    So let’s say if after the 123 constituency seats are filled the numbers are FF 60, FG 40, Labout 14, SF 4, Green 3, PD 2, SP 0 PBP 0, independents 0 then you would move through the parties starting with the smallest PBP as they should have gotten one seat based on the national %, you look through the constituencies and find the PBP person with the high vote and deem them to be elected as one of the 43. That fills their allocation and you move onto the next smallest party and so on. d’hondt works downwards I think, this is working upwards.

  • Drumlin Rock

    does this not create “second class” TDs?
    I think the whole idea of manipulating the vote to suit the minor parties is a tad undemocratic.
    Also as a result of this system all parties would need to stand in all constituencies, and “independants” would have to form parties as well to get those top up votes needed.

  • The idea that some would be 2nd class TDs is one that comes up again and again with whatever top up system is used. My own preference is that a reformed Seanad is the way to get more people without such overbearing constituency ties and for greater proportionality. But I thought it worth outlining a system that addressed the complaints from some quarters that we need more proportionality and that we must avoid a party HQ stranglehold over who gets to run, in addition to addressing the problems of under representation of certain people in the candidate line up.

    Indeed the same argument about being 2nd class representatives is used against the idea of appointing people from outside politics to cabinet (which at present can be done for 2 cabinet members via the Seanad). It’s not done at all and I would sense that the cabinet if drawn mostly from the parliament but with a minority of non-parliamentarians might be a recipe for division.