Is Scottish MMP being “gamed”?

As noted in an earlier comment thread, initially by Dave Hutcheson, former First Minister of Scotland, Alex Salmond, has formed a new party called Alba. It plans to run only list candidates. This has raised some questions about whether this is a “gaming” of the MMP system to enhance the majority of pro-independence parties, bypassing the compensation mechanism. For instance, a pro-independence voter could vote for the Scottish National Party candidate in their single-seat district and the Alba list.

I don’t believe this is true “gaming” in the sense of the dummy lists we have seen in MMP systems in Albania and Lesotho in the past (or even the recent Korean election), but “gaming” does not have a precise definition.

I recommend reading Dave’s comment (linked above) and then the following comment in the same thread by JD Mussel as well as the Politico story he links to. There will also be another new party entering on the pro-union side, All 4 Unity, headed by good old George Galloway (about whom I have written before, most recently under the title, “Galloway is back”; well, he’s back again!). It is similarly motivated to Salmond’s entry: to enhance the total seats for his side of the divide through encouraging tactical split voting.

Then, to get a sense of just how Alba’s entry could affect the result, Leonardo Carella has a very interesting and valuable Twitter thread, viewable in one page thanks to Thread Reader, with simulations under various scenarios.

As I said, I have my doubts this party entry is as problematic as some see it, but it is a debatable point. So, what do readers think?

65 thoughts on “Is Scottish MMP being “gamed”?

  1. Interesting! You can add Venezuela in the 2000s to the list of countries where parties gamed the system (i.e. the compensatory tier) by running entirely made up lists with candidates from the majority party. In 2010, I believe, Chavez officially changed the system from MMP to MMM.


    • I knew it had been changed to MMM, but I did not know it was another case where it had been “gamed” while still MMP. (Hardly a surprise, though, given Chavez.)


      • Don’t blame Chavez for starting it. Yaracuy governor Eduardo Lapi (a Chavez opponent) was quite clever in 2000 in bypassing MMP: he founded a new party to support him. His original party, Convergencia, ran only on the state wide list. The new party was L.A.P.Y. which improbable and coincidental name meant Lo Alcanzado Por Yaracuy (“what Yaracuy reached”, as a clear allusion to how good a governor Lapi thought he was). LAPY ran only on the individual districts. Thus in 2000 LAPY got the 4 individual seats. Convergencia on the state wide list got 2 out of 3 and thus Lapi ruled his term with a very comfortable 6 to 1 majority.


  2. To be honest, I think it would be pretty hard to “game” the system in this way enough to significantly distort the outcome. It doesn’t sound wholly dissimilar to the way the NZ Greens campaign, and I imagine Salmond may find the experience similar.

    In their experience, they tend to get a better party vote where they stand a local candidate. So running list-only would result in a drop in the overall list share. For a party with a celebrity politician like Salmond at the helm I imagine that might be avoided somewhat because they’ll have profile without trying. But, that’s where voter perception will kick in.

    Without running local candidates, and with someone very associated with the SNP leading, a lot of voters will probably feel cheated. They don’t tend to like that, so he may need to field local candidates to gain legitimacy as a party.

    But, having worked through what you would need to do to distort the system this way here in NZ I think I see the opportunity Alba is meant to exploit – the SNP has such a tight grip on electorate seats that any new addition could create a distortion, it will just be a matter of Salmond’s ability to convince the electorate he isn’t cheating, and how many bridges he’s burned with the SNP by creating a competitor.


    • I think it’s worth pointing out how explicitly Alba is presenting their strategy. Scroll down on their website and you’ll quickly see their claims that “almost one of two pro-independence votes go to waste because of the Additional Member System (AMS) used in Scottish elections” and that “the #Supermajority strategy creates the opportunity to secure over one million additional votes for independence” prominently displayed, before any mention of policy: indeed, at present there seems to be basically no mention of policy at all. The Scottish Greens also generally try and get the list votes of SNP voters, but they do run on a distinctive policy platform.


      • “Go to waste because of the Additional Member System.” That’s either ignorant or malicious. In a compensatory system, if a party gets enough local seats it does not deserve additional seats. These Albas are not only trying to sabotage their old party, they are also trying to sabotage their voting system. I hope the media are paying close attention.


      • It’s neither ignorant nor malicious, though it certainly doesn’t care for norms of proportionality. Votes that fail to change the result though they could have done so are “wasted”, by definition. Objecting to the description by saying those votes didn’t “deserve” more representation is rather like saying wasted votes don’t “deserve” any more representation under FPTP, so they’re not really “wasted”.


    • Thanks for this, Ben. I agree completely. In fact, a pretty key effect of (two-vote) MMP is that parties do tend to get better list votes from having a candidate (of some decent quality/fit to the constituency) than if they decline to enter the district contests. I wrote about this–what the literature calls “contamination”–with respect to the NZ Greens in 2014.

      The claim about the importance of district candidates is also supported in the results of my forthcoming coauthored Party Personnel Strategies book, albeit tested only for large parties.

      Regarding the “tight grip” of the SNP on constituencies, it seems to me that if one takes this as “gaming” (and as I’ve said, I mostly do not), it would be less consequential if the Scottish MMP had a more proportional overall design.


  3. As I’ve said before, I think that allowing the runner-up in each district to designate which list the district winner gets “deducted” from, should put a cap on gaming MMP.


    • Since this post is dated April 1st, I have to assume that you’re not serious. It would turn the immediate post-election period into a game show, or something akin to the ‘Club Call’ system used in the rugby-league play-offs from 2009 to 2014.

      It would also be a surefire way to turn voters against MMP if they observe how “losers” (the runners-up) in each district can direct their votes to whomever they want for tactical purposes. It’s like the preference flows at the Western Australia Legislative Council election, only somehow you’ve managed to make it even worse.

      The best way to stop gaming MMP is with the single-vote variant, which we saw in action in Germany just three weeks ago.


      • See my clarification below – this should apply only to district winners not affiliated with a top-up-seats-winning list and I ought to have made that clearer the first time.


      • The single-vote variant would be worse. As Enid Lakeman pointed out when the Hansard Society proposed it, small parties would stand candidates in every district just to make sure they reached the 5% threshold. And the big attraction of MMP is usually “You get two separate votes, and can re-elect a hard-working local MP while supporting the party you want to win.” Supposedly, we were told, parties listened to voters so if a local candidate polled well on first-votes, they would get moved up the list, while a sitting MP who polled unexpectedly poorly in their district contest, they would be demoted in the party ranking. (The NZ Royal Commission sold this feature as a way of ensuring some individual accountability for MPs, without going so far as to allow voters to choose among competing candidates of the same party, because the latter would undermine democracy, as it has done in Austria, Belgium, Brazil, Chile, Denmark, Finland, Ireland, Malta, Sweden, and Switzerland.)
        But if and once it’s rejigged to “single-vote closed lists” – so that you have to vote for an unsatisfactory local candidate if you want to support that party at all, like FPTP, but with the added advantage that, unlike FPTP, you may not even be able to keep that local candidate out of the next parliament and have no way to signal to the party machine that you’re dissatisfied with their choice of personnel… this selling-point disappears.
        I’m not against parallel PR systems, especially since then the parallel PR seats can be filled via STV instead of lists (not possible with compensatory PR as far as I can tell), but if one wants to keep MMP, maybe a compromise would be to keep two votes and add each party’s totals of list and candidate-votes together (not necessary in a 50-50 ratio) to allocate seats?


      • Even single-vote MMP could be gamed. A party runs its ‘A’ candidates in districts where it thinks it will win and its ‘B’ candidates where it expects to lose.

        However, the potential reward is much smaller than with two-vote MMP and calls for good polling info. A risk is that the elected B members may begin to act as though they truly were an independent party. Such a reward/risk ratio almost certainly would be too low to find any takers.


      • Most MMP ballots I’ve seen, whether from Dresden, Dunedin, or (er) Dunedin, have about 50% more options on the party-list side than on the candidate side. Say, 7-8 candidates and 10-12 (or more) lists.
        As noted, Enid Lakeman observed that small parties, if unable to draw votes by standing a nationwide or region-wide list, would have to draw votes by standing a candidate in every district. Trust me, speaking as an Australian you do not want electoral rules that encourage minor parties to stand candidates just to get ballot presence. Bad idea all round! Let them run five or six candidates on a national/ regional list, if they can only hope to win one or two seats, and don’t force them to run a candidate in each of ten or thirty or sixty districts before they can receive any votes at all.
        With more minor-party candidates in each district (who have no chance at that level), election of a local MP will become even more chaotic than under “vanilla” FPTP (since the incentive to “keep out the lesser evil locally” will this time collide head-on with “support your first-choice party at national or regional level” and it is not clear which incentive will give way). MPs may be elected with 20% or 25% rather than 40% or 45%. A Neil Hamilton may squeak back in and then claim a spurious mandate because the Labour and Liberal Democrat parties did not want to stand down for a Martin Bell. When Labour and LibDem stood down for Bell in 1997, it did depress their national vote totals slightly, but this was symbolic only and did not cost them any seats: indeed, replacing Hamilton with Bell counted as a half-point for them. But with one-vote MMP, their standing down in a Hamilton-vs-Bell type contest would definitely cost such parties top-up seats.


  4. Having a district runner-up designate the party affiliation of the district winner would give major parties an interesting new tool in their fight to suppress small parties.

    Imagine rerunning a UK election a dozen years ago, with a Conservative runner-up able to say that the Labour district win should be subtracted from UKIP’s list or a Labour runner-up asking for a Tory win to be subtracted from the SNP list.


    • Sorry, I’m missing something. Why exactly would a Tory candidate want Labour to win (say) 249 seats instead of 300 in order that UKIP wins only 9 seats instead of 10, by deducting a Nigel Farage win from Labour’s column? Are the Conservatives more likely to ally in that next Parliament with Labour than UKIP? Sure, they don’t want new rivals supplanting them on the Right, but wouldn’t putting together a majority right here and now be more pressing than the risk that UKIP might grow like 1920s Labour or 2000s SNP rather than wither on the vine like the SDP?


      • Apologies, typo – should say “want Labour to win (say) 300 seats instead of 249”, but my query stands. Why would it be a stronger incentive to suppress the hypothetical future growth of a rival party in future parliaments than to immediately cobble together a majority against the main opponent in the current parliament?
        I’m not seeing why Labour 300 MPs (overall), UKIP 9 has greater attraction for the Tories than Labour 299, UKIP 10.
        The behaviour of UK politicians putting together post-election alliances – the Lib-Lab Pact in 1974, the Conservative/ LibDem Coalition in 2010, the DUP supply-and-confidence agreement in 2017 – shows they may need every seat their frenemies hold to stop their enemies. Wilson didn’t care that his overtures to Jeremy Thorpe were construed by the Liberals as the new dawn of their party’s revival, he just wanted to block Heath. (In any event, Thorpe own-goaled that revival very dramatically).
        I ought to clarify that allowing, say, a defeated Tory runner-up to tell the Electoral Commission whether Nigel Farage’s win belongs in the “Labour” or the “UKIP” column should only apply if the district winner does not appear on any list and is not endorsed by any party – ie, independents or very small parties. (The German solution is to throw out all list votes cast by a winning independent’s supporters. Arguably this should only be done if they number less than a Hare or Droop quota of the entire parliament, otherwise should be reduced to a surplus fractional value, as with SDP. Of course with 496 Bundestag seats, polling over 1/496th or 1/497th of the nationwide votes would usually require an absolute majority in one of 248 districts).
        Arguably the “runner-up decides” rule could also apply if the district winner was endorsed by a party, and/or was named on a list,* but one that fell below the 5% thresholdd. Eg, in NZ, if ACT-NZ wins only one or two districts, the runner-up (presumably Labour) could then state that these are billed to National. Of course this might then backfire if ACT-NZ were to do a Winston Peters and back Labour instead (its founder was, after all, a former Labour Treasurer).
        Someone, not here, once suggested that an unpledged district MP could be deducted from whichever list polled the most Zweitstimme in that district, but then this could be gamed or swayed by the vagaries of FPTP. Eg, Red’s candidate wins the district 45-40-15% on Erstestimme, because of a strong third-placed Light Blue candidate, but Dark Blue wins the Zweitestimme (without the Light Blue spoiler) 55-45%, so Red keeps the district seat but Dark Blue gets “billed” for it.


      • Aaaargh another typo while correcting a typo. “reduced to a surplus fractional value, as with SDP.” ==> “reduced to a surplus fractional value, as with STV”. Doesn’t or didn’t Hungary do this?


  5. Agreed, a party would not opt to throw away a chance to form a government by such a tactic. I was thinking of the situation where it’s clear that the party has no way of forming a government — say a long-in-the-tooth governing party loses the election with an upstart on its unprotected flank. No point in reducing the upstart from 10 to 9 seats, but it might be tempting to take them from 5 to 0 to deprive them of a parliamentary platform.


    • I see what you mean. They could do that, but there would be costs in terms of

      (a) public backlash against a case of – for once – clear and blatant electoral manipulation (unlike the usual snoozeramas over “replacing Harmonic Mean with Hagenbach-Bischoff” that fascinate all of us here but lose the average voter), and

      (b) starting a long-term vendetta against a rival on the same side of the ideological fence whose MPs they could well need and whose voters they might otherwise have won over.

      It depends whether the cost of this trick is less overall, in frequency and/or severity, than the cost of allowing decoy lists. (“No, the candidates who swept 16 out of Queensland’s 20 federal district seats are all from the NATIONAL Party, which polled 0% of list votes. The LIBERAL Party, which polled 52% of list votes, hasn’t won any seats at all yet, so it’s owed at least 6 of the 11 top-up seats…”)


      • Yes, I suspect some of the most pernicious examples of parties gaming Tom’s system (Australian Labor asking for Liberal winner votes to be taken off the Greens total, for example) would be helped by a need for minor and major parties to have reasonably good long-term relationships under a proportional electoral system, although I am somewhat hesitant that this approach could work in Tasmania. Absolute majorities are rare under MMP anyway, so I think under most circumstances it would be worth the while of even an unpopular losing party to try and reduce the seat count of the winning party (for example, in a 100-member legislature, a losing major party might be indifferent between the winning major party having 55 and 60 seats, but not between 45 and 40, since they could aspire to have some influence under the latter system).


  6. I agree with what Tom says about the disadvantages of single-vote MMP (03/04/2021 at 6:10 pm) and also with what Dave says a little later about how even the single-vote version can be gamed.

    See the post by Manuel on the 2018 Mexican election, for instance.

    (Mexico does not use MMP, but the point still applies. It is actually MMM, but has a partial compensation feature.)


  7. My take on a twist of Jenkins’ failed AV+ has made me come up with an improved system.

    Voters get 2 votes: one for a district candidate and one for an at large candidate. Votes cast for both candidates are added up by district. A party getting a majority of the first+second votes gets the district seat. If not, the party with the fewest votes is eliminated and their votes are transferred.

    After the district seats are filled up, the at large seats are allocated. Again, the first+second votes are tallied within the top up district but all votes that went to an elected candidate are gone. This would be the real plus where your 2 votes count towards electing 2 members, instead of 1 or none.

    I came up with this after watching the former binominal system worked in Chile. But the idea of adding both votes, for both tiers, seems to be something radical.

    Adding up first and second votes would do the trick. Voters would rank both ballots.


    • Under AV a voter’s 2nd choice is given as much weight as their 1st choice. Under FPTP the voter’s 2nd choice is given zero weight. It seems that Derek’s proposal allows the voter to make this all-or-nothing decision separately for each of the district candidates. When deciding whether to rank a district candidate, the voter must ask “If this person is elected, is it worth the price of having my at-large vote discarded?”

      What happens with Independents? Can one person run as both a district candidate and an at-large candidate?


  8. First of all, apologies to Tom for saying that he posted on April Fool’s Day. Those pesky time zones!

    I understand why an Aussie would have a particular aversion to encouraging minor parties to stand everywhere as there are probably still recurring nightmares about the Senate ballot paper from NSW all those years ago. But we are not talking about adopting a preferential system here, so a slightly longer ballot paper in most constituencies isn’t something to be feared. I am not against offering voters greater democratic choice.

    In the UK there would still be the £500 deposit and 10-signature requirement to deter truly spurious candidates. (We still have 16 candidates in the forthcoming Hartlepool by-election, and districts with the number of candidates in double figures for actual two-vote MMP elections in Wales and Scotland next month, even with those rules in place.)

    The fact that you have to point to a specific example from one seat in 1997 with unique circumstances – one I remember well – to try to prove why single-vote MMP would supposedly not work or be just as prone to being rorted shows that the argument is weak. The Lib Dems have since stood down in favour of Plaid Cymru or the Greens in particular seats as recently as 2019 anyway and could perhaps could do with a nudge to remind them of the importance of a full slate across Great Britain.

    I wasn’t aiming to defend the systems “in Austria, Belgium, Brazil, Chile, Denmark, Finland, Ireland, Malta, Sweden, and Switzerland” – I was sticking up for the system used in Baden-Württemberg. I do think that the variant used nextdoor in Bavaria would address Wilf Day’s specific concerns, but I’m alarmed at the obstacles people keep finding to the adoption of single-vote MMP rather than seeing its advantages and how simple it would be to implement in countries accustomed to FPTP…


    • Even as an STV shill I could live with a German State-level version of MMP with features like open (or at least flexible) lists; large regions (20-40) instead of nationwide (and perhaps a threshold more like 2% within a region than 5% nationwide); and adding together First- and Second-Votes to get the list totals, as a solid second preference.
      I think this puts me in Wilf’s caravan?


      • I’m reporting you to the central committee for appropriate discipline. Expect to be visited by a large ugly man carrying a violin-shaped case in the immediate future.


      • Alan, don’t forget that STV was not only advocated for the UK by Jeremy Thorpe but actually introduced in NSW (albeit only briefly, 1920-1927) by TJ Ley, the alleged “Minister for Murder”:

        There must also be enough Sinn Fein hardmen on the pro-STV side to give Mussolini, Robert Mugabe, the Orange Order volunteers, Jack Lang’s machine, and the Tammany Hall boys in Team Anti-PR a good rumble.


      • Idle threats aside, I do think that one’s ranking of voting systems should bear some relationship to one’s preferred voting system. My first choice is STV, which means I am prepared to accept a second-best alternative if the best is for some reason (sound or unsound) not on the table. This is not being inconsistent. At some point down the preference-ranking scale, as the utility scores decrease, it becomes better bet to exhaust, truncate or stay home.
        I suppose is also makes sense for FPTP supporters to plump or bullet solely for FPTP and to reject every other voting system (“STV and MMP are too proportional, so they’re bad! – Alternative Vote is not proportional enough, so it’s bad too! Everywhere in the world except where some outpost of white colonists has to live under majority rule by the native population, FPTP is the only acceptable voting system!”).
        The ones I find really inconsistent are the Approval diehards. By their own logic, AV-STV should receive either an equal tick with Approval if it’s an adequate improvement over FPTP, or it should receive no support whatsoever if it’s not as good as Approval. Instead, the usual stance of Approvalistas – grudging recognition that “well, at least AV-STV keeps out Condorcet losers, but it doesn’t guarantee the election of Condorcet winners the way that Approval would do” – makes no sense within their own terms of reference. Tick or no tick, Approvalistas: there is no gray.


        • I don’t agree with this take on Approval voting supporters. I prefer non-ordinal procedures largely for reasons given by Arrow and May. So I take Approval to be superior to RCV for single-winner elections (only). But for multi-win elections I don’t think it’s much good and prefer SNTV or list systems. But I’d take STV over any non-proportional plurality winner mechanism. It’s all in my book :

          Anyhow, the idea that all approval vote supporters are the same is silly.


    • Tougis – apologies if I’ve misrepresented you. I’m judging Approval by its most vocal supporters, both those who periodically show up at F&V and those across the Internet generally. And they all seem to detest STV-AV.
      Myself, I think Approval would operate in practice – above the level of non-partisan local elections – as a very rough and ready form of AV. Greens voters would tick the Democrat as well as their own candidate; Democrats would have no incentive I can discern to tick the Greens candidate as well as their own nominee; so a result of Greens 12%, Democrat 47%, GOP 43% on supposedly multiple equal first preferences would approximate a result of Greens 12%, Democrat 35%, GOP 43% -> Democrat 47%, GOP 43% under IRO-AV.
      Oliver – most of the complaints about the larger size of STV electoral districts seem to overlook that each district will have three, five or seven representatives, not just one, and that if voters really do prefer to elect candidates who live close to them – as FPTP and MMP supporters alike assure us they do – then a candidate in an STV district need only canvass for votes in the northern or the eastern third or fifth or seventh of their district, just as they do with SMDs.
      If that is not in fact the case – if STV would indeed require the former MP for Oxford South-West to go chasing votes across the entirety of Oxford to have any hope of election – this seems to me to concede, in a way that does not help the case for FPTP and MMP, that there are currently voters in northern, central and western Oxford who are being denied their first choice of representative by the current single-member electoral boundaries. True, Ireland and Malta are small in area and population, but if the UK can have Commons constituencies with 100,000 voters and Australia can have House divisions with 150,000, the problem is scaleable.
      Moreover, it would no longer be possible to gerrymander single-seat districts within that region, and voters would not be corralled into supporting local candidates by lines on a map, but I count that no great loss.


      • I live in the largest STV district in the world. NSW is a single district for the 12 federal senators and the 42 members of the NSW legislative council. The area of the state is 809 952 km2. The population is 8 166 369. I live in a town of less than 20 000 that is over 1100 kilometres from Sydney.

        If there is anyone here with some claim to be an oppressed victim of remoteness and sparse population it’s me. Much as I love my neighbours they tend to be just a tad conservative and I would be incredibly unlikely to approach the local lower house members about any issue that involves me. Instead I have to round up my bullock team, get them into traces, and undertake the 3 month journey to Sydney by bullock cart so I can talk to a senator or MLC.

        It’s quite a difficult journey. As the poet Judith Wright wrote in Bullocky:

        Till the long solitary tracks
        etched deeper with each lurching load
        were populous before his eyes,
        and fiends and angels used his road.

        I am thinking seriously of getting a phone installed but its seems a bit newfangled to me.


      • Alan, I would like to see figures from how many voters do in fact talk to “their” “local” MP face to face as opposed to using telephone, Skype/ Zoom, social media, and other such technologies.
        I suspect the proportion is probably much higher for National Party supporters, who of course cluster in rural areas, probably because people who vote National are more conservative by temperament, and also because so much of National Party policy-making is transactional rather than ideological – you rail against “socialism”, of course, but you also want a new government-provided road in your area, so you want a quiet word with Gary Kotter III, your local MHR. You did his grandfather, Gary Kotter Senior, a solid with some stock deals back in 1966 so Trey has inherited that debt and owes you a favour. If your district ever elected a different MP, those debts would vanish like Alan Dean Foster’s Star Wars novelisation royalties, so you want to keep the seat in the Kotter clan as long as possible.
        In addition, my suspicion is that a lot of National Party political negotiation involves matters that you don’t want to be discussing by phone, internet, or any other artificial method that the Independent Commission Against Corruption might be able to tap. It won’t be about Rawlsian vs Nozickean conceptions of distributive justice. Lest I be thought cynical, the NSW Nationals leader, John Barilaro, proudly claimed the title of “Mr Pork-Barilaro” a few months back.
        So yes, there may be significance, perhaps, in the fact that the world’s sixth-largest country (and fourth-largest competitive democracy) in land area consistently adopts STV-PR and rejects party lists?


  9. I’m not sure I’d call myself an “Approvalista”, but the more positive type of approval voting (i.e. ‘put an X next to all the candidates you could stomach as your MP, rather than ‘Yes’ and ‘No’ boxes next to each candidate) is probably my next favourite alternative to FPTP after single-vote MMP.

    The advantages in terms of convincing people to make the switch are clear:

    1) For parties and candidates, the way of nominating candidates would be the same as at present;
    2) For voters, the ballot paper would look exactly the same as at present, including the instructions for how to cast a valid vote; and
    3) For tellers examining the votes on a constituency-by-constituency basis, the method for counting and adding up vote totals for each candidate would be the same as at present.

    The only officials with a bit of extra work to do would be the regional returning officers. If, after a few cycles under that system, people decide that they want two votes instead of one, and/or that there should be lists, then those further reforms could still be made.

    As for STV, I’d be prepared to try it out in local elections in England, but the experience of the change to that system in Scotland serves as a cautionary tale. STV for either chamber of the UK Parliament, on the other hand, would not be suitable at all. Constituencies with even the minimum level of proportionality (3 members to be elected) would have populations of over a quarter of a million, which is in stark contrast to the genuine localism of Ireland, Malta and Tasmania. And I don’t want a 5,000-member House of Commons either, thanks.


      • With respect, neither you not Tom has convinced me. (Especially since ‘Oxford South West’ would be far too small for even a single-member constituency!)

        I don’t really count the variants on STV used for the upper houses of NSW, SA, Victoria and WA as ‘true’ STV. They’re really forms of list PR adapted to the constraints of a peculiar misreading of the Australian Constitution by that country’s High Court. We currently have a legal maximum area for a constituency of 13,000 sqkm and – although it must sound absurdly tiny to a rural Aussie – I don’t think there’s scope for going much larger than that. A 3-member district for the Scottish Highlands and Islands could well reach more than double the current limit.

        Under no-list MMP you’d continue to have both a local representative and multiple regional members from across the political spectrum whom you could contact with your concerns. Whichever candidate(s) ran in your division and took an additional seat could even be seen as an alternative port of call, as already happens under two-vote MMP in New Zealand. And in your state, you’d still have two chambers so a different system would be used for whichever one didn’t adopt MMP. You have nothing to fear from such a move.


        • The high court has never held anything with respect to the composition of sate upper houses. Not only have they not misread the constitution in this matter, they have nothing to misread because the federal constitution does not address the topic. Moreover, your definition fails for the Hare-Clark jurisdictions, Tasmania and the ACT, because they do not use any form of GVT and for New South Wales and South Australia which have fairly generous provisions for voting below the line voting.

          With the greatest possible aspect you have offered a distinction without a difference based on high court rulings that do not actually exist.


        • JD

          The MLC I usually deal with lives in Sydney’s Inner West about 10 minutes from Parliament House. The only district with representation in the NSW legislative council is coterminous with the state and I am more than happy with the constituency services and political advocacy I get from ‘my’ representative even though she lives 1143 kilometres away. And I don’t communicate with her by bullock train.

          In any district-based system I would almost certainly be dealing with an MLC with whom I would have zero, zip, nada in common beyond the accident of geography.

          I think it’s interesting that only 2 of the 23 full democracies according to the Democracy Index—Germany and New Zealand—use MMP and neither have an elected upper house. Ii”m ready to be corrected on that, this is back of the envelope stuff) Electoral commissions in Australia already have quite enough difficulty with electoral education explaining why people receive lower and upper house ballot papers with slightly different voting rules. If, say, NSW adopted two-vote MMP for both the assembly and the council, explaining why people were receiving 4 ballot papers might become a bit of a challenge. If the assembly went MMP and the council stayed STV explaining why they had to vote numbers on one paper and a single X on the other would be even more challenging.


      • Alan

        What you say makes sense to me.

        I was responding to Oliver’s advocacy of ‘best losers’ single-vote MMP (a type of district-ordered list system).


  10. What Alan says is, of course, technically correct. My explanation took a few shortcuts. The Australian Constitution doesn’t mention state electoral systems and the High Court has only ever ruled, to my knowledge, on federal elections. The ruling in question concerned “direct election” and the Court took that to mean “an election implying a personal mandate” rather than “an election by the people, as opposed to indirect election via the state legislature or an electoral college”. This is why preferential systems have become part of Aussie political culture, which in turn heavily influences the choices of the states and territories when deciding on electoral systems.

    I did not claim Tasmania and the ACT failed at having ‘true’ STV. They definitely fall under my definition. They have a population of 431,000 and 541,000 respectively. The UK has 67.6 million. STV would only be suitable for local elections at best. Single-vote no-list MMP remains the way to go, at least to start with.

    The lack of group ticket voting in NSW and SA is welcome, but the savings provisions (whilst good at stopping spoilt ballots) bring the system in effect closer to list PR, in my opinion. Your kilometrage may vary.

    The point about different instructions for different ballot papers is an interesting one, and I think it (mostly) strengthens my argument – albeit with a caveat for bicameral jurisdictions where both chambers are elected simultaneously via different systems.

    In 1999, at the first elections to the National Assembly for Wales, it was said that Plaid Cymru did especially well because Labour voters didn’t realise they could give both votes to Labour. Single-vote MMP (even a variant with lists) would’ve spared that confusion. In 2007, STV was introduced for local authorities in Scotland and elections took place on the same day as those to the Scottish Parliament. There was a record number of rejected ballot papers as some voters wrote numbers on the Holyrood ballots. Since then, local and parliamentary elections have no longer been held concurrently.

    Next month in London, voters will be faced with 3 ballot papers (or even more if there’s a borough council by-election happening). On the first, they can mark up to two preferences for mayor… but with an ‘X’ in separate columns, because the government thinks the English are too dim to understand figures. On the second, they get one vote for their ‘local’ Assembly member (where the constituencies have an average population size of 650,000!) and on the third, another ‘X’ for their favoured party. It’s an absolute mess of a mix of ideas. And could certainly be improved with some simplification.


    • I am afraid I am still unable to recognise any high court decision from your quotations and I am forced to ask for a citation.

      It is not a tenable argument that the savings provisions make STV into not-STV because (1) identical savings provisions exist in the ACT and Tasmania and (2) all other STV systems make the expression of preferences optional. It would follow as a matter of logic that it is not only the state upper houses, Tasmania and the ACT, but all STV systems anywhere on the planet that are actually not-STV systems.

      In any case, the number of people in Australian STV systems, or not-STV systems, who vote only one preference is very small. I’ve chosen to exhaust my vote a couple of times, not because I was secretly voting FPTP but because I didn’t want to vote for the remaining candidates. I also on one occasion, feeling eccentric, filled out all 257 individual squares on a NSW legislative council ballot.


      • Alan, Oliver clearly altered his argument (or caveated it). Above, you did not deny there having been a High Court ruling on the federal electoral system, only regarding the states. Are you now denying a federal ruling as well?


      • There have been a number of decisions about the federal electoral system. None, as far as I know, have mandated preferential voting in the way that Oliver claims. That is why I asked for a citation. Before I made that request I carefully reviewed <a=href “”>Day v Australian Electoral Officer, which addressed the validity of the 2016 amendments to the Commonwealth Electoral Act 2018. This article by Anne Twomey analyses Day’s arguments in better language than the court.

        In Oliver’s comment he wrote:

        They’re really forms of list PR adapted to the constraints of a peculiar misreading of the Australian Constitution by that country’s High Court.


        There is no mention in the decision in Day’s case, in any decision cited in that decision, or in Twomey’s article, of a constitutional requirement for preferential voting. I am aware of no decision that imposes a constitutional requirement for preferential voting. Oliver cites language in quotation marks which indicate that he is quoting the court. It is perfectly reasonable to ask for the relevant citation.

        I have already addressed his contention that above the line voting, or optional preferential voting, mean that STV is not STV.


  11. I’m afraid I don’t have a citation to hand, but it’s something I recall reading many years ago in the comments on Antony Green’s old blog. It certainly isn’t as recent as the Day case. Nor did it mandate preferential voting, but certainly appeared to rule out any sort of system – such as closed-list PR – whereby an elected politician cannot claim a personal mandate.

    As for savings provisions, perhaps they indeed are not the problem. It’s plain to see that the Irish and Maltese variants of STV are ‘proper’ STV, despite the number of preferences being optional. Perhaps it’s to do with ballot design? I know Ireland does not group candidates by party. That definitely pulls it further away from resembling list PR.


    • The idea of grouping candidates was first proposed by Enid Lakeman. Candidates are grouped in Tasmania and the ACT, although they also apply Robson Rotation. It follows that if it is grouping candidates, and not as you first proposed GVTs, that make STV into not STV, then Tasmania and the ACT do not have STV either.

      To summarise:

      There is no misreading of the constitution which mandates preferential voting, GVTs, candidate grouping or anything else
      OPV does not make STV into not STV
      Each of the three successive tests you have proposed that make STV into not STV applies in Tasmania and the ACT, and 1 of them applies in all implementations of STV outside Australia

      Despite grouping candidates, in Tasmania the parties generally no longer recommend a ticket order because Robson Rotation means there is no way they can enforce it. I do not know if that is also the case in the ACT. Tom knows the ACT better than I do. Tasmania has elected senate candidates, under the more restrictive federal system, against the recommended ticket order.


  12. “There is no misreading of the constitution which mandates preferential voting, GVTs, candidate grouping or anything else”
    True. There is, nonetheless, a misinterpretation that would prevent a (d’Hondt/Sainte-Laguë) PR element from being adopted federally. The High Court’s understanding of ‘direct election’ wouldn’t rule out a reversion to FPTP, for instance, so it clearly doesn’t mandate a preferential system and I never said that it did so. I merely stated that it strongly nudged Aussie political culture in that direction.

    “OPV does not make STV into not STV”
    Naturally, since Ireland and Malta have ‘purer’ STV than most Aussie jurisdictions. I’m not sure how you twisted my words about GVTs into that confusion in the first place.

    So we now have a hierarchy of factors in determining how ‘proper’ a claimed form of STV is:

    -No GVTs (pretty crucial for it not to be pseudo-STV) – two states Down Under fail this test
    -Size of jurisdiction by population and area (feels more like ‘true’ STV if these are small and candidates can get to know the electorate well and vice versa) – only Tasmania and Canberra can realistically pass this one
    -Ballot paper design (mild effect in influencing voters to think in partisan rather than individual terms; probably disadvantages independent candidates) – am grateful for your research on this one; Robson Rotation would be welcome under any system in my opinion, although it can make manual counting more complicated for tellers
    -Optional or compulsory preferential voting (immaterial) – some states and territories seem to go back and forth on this one anyway

    All of these arguments could be avoided if Australia would just ditch its addiction to preferential systems for both houses. You can always keep them for one chamber if you’re that attached! [All of this has become quite a distraction from discussing whether the Scottish election is being gamed by Alba. Sorry about that.]


    • Australia is not about to ditch the alleged addiction to preferential voting because of a series of blog comments based, among other things, on high court opinions that do not actually exist.

      Your hierarchy of tests for not STV merely reiterates a series of factual claims that have already been demonstrated to be partly or wholly wrong.


      • Olivier would be thinking of the 1984 McKenzie case where the High Court held that ticket-voting was constitutionally acceptable provided it was only a quick short-cut to voting for individual candidates. Any electoral system that prevented voters supporting candidates individually would, in Australian constitutional law, run afoul of the federal Constitution’s requirement that Senators (and MHRs) be “directly chosen by the people”. Closed party lists would certainly be unconstitutional for Australian federal elections and it is likely that even flexible, free or open lists would be invalid, on the reasoning that even these deprive the voter of their right to decide their own combination (if FPTP or MNTV) or permutation (if STV or IRO-AV) of favoured candidates. (Ie, even with a Finnish open-list system, if you vote for Bernie on the Democrat ticket, your vote could help elect Joe Manchin instead; had you known that, you might have voted for the Greens ticket with Nader, and there would be nothing you could do to stop that short of not voting for Bernie…).

        This requirement does not legally bind the State Parliaments for State elections. The nearest rule that does bind the State Parliaments for State elections is an entrenched provision, first in the Colonial Laws Validity Act 1867 and now continued by the Australia Acts 1985-86, that only a “representative” legislature enjoyed plenary law-making power, including the right to alter or repeal English common law as transplanted to the Empire’s colonies. A “representative” legislature was defined as one of which “at least One Half” of the members were elected. (Ironic to think the British Parliament would not count as “representative” under this definition! Like “the more numerous house”, it runs into problems when the Lords Spiritual and Temporal outnumber the Commons…).

        The NSW Parliament had an indirectly elected upper house as recently as 1978. (Every 3 years, 15 of 60 MLCs were elected using PR-STV by a joint sitting of the 90+ MLAs and the 45 non-retiring MLCs). Queensland and NSW had appointed, life-termed Legislative Councils until 1922 and 1932 respectively. No one ever suggested these were invalidated by the 1901 Constitution.

        (The closest anyone came to making an argument that the federal Constitution regulated the structure of State Parliaments was that sec 15, which requires Senate vacancies to be filled by “both houses of the State Parliament” voting as a joint sitting, forbade unicameralism, but the High Court rejected this claim).

        So, States could legislate for party lists (and have in fact done so, viz South Australia’s Legislative Council 1973-1984). The Commonwealth can legislate for party lists for bodies other than the House and Senate (and has in fact done so, viz the ACT’s Legislative Assembly 1989-95). It will be observed that both those experiments were quickly terminated.

        Now it is true that STV as practiced for the mainland upper houses (federal Senate, and the NSW, WA, SA and Victorian upper houses) is more like a list system in practice than the “Hare-Clark” version used for the Tasmanian and ACT lower houses. These systems were most “list-like” in the 1980s and 1990s when used only for the first four upper houses mentioned; all combined with “one in the box” full preferencing with (apart from NSW) mandatory numbering of all candidates as the sole alternative.

        However, reforms in the last 15-20 years have moved away from this highly regimented model. In 2003, NSW replaced “a 1 numbers all candidates” with preferential voting among team groups – a reform copied by the Commonwealth in 2016 and South Australia last year. And when Victoria adopted PR for its upper house in 2004, it required only a minimum five preferences for individual candidates. So WA is the only remaining Jurassic Park with the “number every candidate” rule.

        To reiterate, the Australian States were never legally bound to copy the federal Senate electoral system. Rather, they followed it voluntarily, primarily as a “best practice” example, but also because the drive for uniformity between federal and State electoral rules is often a tiebreaker – different rules mean more spoiled and invalid votes – and the States are more likely to blink first in that game of chicken.


      • Olivier cannot have been thinking of Mackenzie’s case. In his first comment Olivier wrote:

        I don’t really count the variants on STV used for the upper houses of NSW, SA, Victoria and WA as ‘true’ STV. They’re really forms of list PR adapted to the constraints of a peculiar misreading of the Australian Constitution by that country’s High Court.

        McKenzie’s case is cited in Day’s case and is therefore one of those I reviewed. It says nothing to mandate GVTs, and while the high court has unusually wide jurisdiction for an apex court, it is not vested with jurisdiction to make orders in 1984 that command the legislature to make particular laws in 1918, 1948 or even 1983.

        The ratio decidendi in the case is:

        6. The question that now falls for decision is whether the provisions of the Act to which I have referred are open to objection on constitutional grounds. The plaintiff submitted, first, that electors who use the simplified system of voting will be voting for parties and not for candidates and that this will contravene s.16 of the Constitution which provides for the qualifications of a senator: it is right to say that the electors voting at a Senate election must vote for the individual candidates whom they wish to choose as senators but it is not right to say that the Constitution forbids the use of a system which enables the elector to vote for the individual candidates by reference to a group or ticket. Members of Parliament were organized in political parties long before the Constitution was adopted and there is no reason to imply an inhibition on the use of a method of voting which recognizes political realities provided that the Constitution itself does not contain any indication that such a method is forbidden. No such indication, relevant to the present case, appears in the Constitution.

        There is no reference to any kind of list in that paragraph or anywhere else in the judgment, and as already conceded, neither the constitution nor any judgment based on it imposes particular electoral systems on the states.


    • Thanks for bringing this to our attention, and also to Tom for confirming I wasn’t going completely doolally by imagining an Aussie High Court ruling. I am particularly tickled by the additional nugget of information that the UK Parliament would fail its own test for ‘representativeness’!


  13. With 32 of 73 constituencies reporting both constituency and list vote results for Thursday’s Scottish Parliament election, Alba is doing poorly on the list vote, with just 1.7% so far, while Galloway’s AFU is even worse off at 0.9%. Prior to the election I was highly skeptical of claims about Scotland’s MMP system being gamed, but these results make it clear that if somehow that was the actual intent, it appears to be going nowhere at this stage.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.