No-one can be in any doubt about my attitude towards group voting tickets – they need to go, and in the meantime, everyone should vote below the line in Victoria.
Often the arguments in favour of group voting tickets often focus on diversity in election – getting ‘the little guy’ elected and having some diversity in the system. This ignores the potential for corruption caused by GVTs and how it hinders the growth and sustainability of minor parties. It also lumps together all of the people who vote for minor parties while flattening the range of opinions they hold. It doesn’t matter who is on the crossbench as long as they’re interesting and diverse.
But it is also true that the current Victorian electoral system would produce less political diversity in the absence of group voting tickets. By my estimate, the current regions would have elected four extra Greens, an extra Shooters member, three extra Coalition members and one extra Labor member, with six other parties wiped out.
This is in some ways a fair result. It eliminates a lot of the wheeling and dealing of group voting tickets and it favours a handful of strong minor parties over a large number of weak minor parties. But I think you can make a case that it is preferable for the barrier to entry to be lower.
I spent a lot of time last year banging the drum about the importance of magnitude (the number of representatives elected per constituency) in proportionality. An M5 district is not that proportional. It certainly won’t represent parties getting a few percentage points. That might be ideal in a house of government, where you want a relatively small number of parties to aid in formation of stable government, but in an upper house you can be more relaxed.
I think the answer to this dilemma is to pair the abolition of group voting tickets with the abolition of the regions, electing all 40 seats in the Legislative Council at large statewide with a quota of approximately 2.4%.
This chart compares the actual 2018 result to the result without GVTs, and also makes an attempt at estimating the result if the chamber was elected at large (M40). For this purpose I’ve ignored preferences – the experience of NSW suggests preferences could change 1-2 seats but they would be far less important with the much lower quota.
The abolition of GVTs mostly boosts the Greens and the Coalition. Statewide election, however, reduces the number of major party MLCs (and to a lesser extent the Greens). It would have given two seats to Hinch's Justice Party and one each to six other parties. Four of these parties were the same that won seats in reality, but the DLP and Voluntary Euthanasia would have won seats instead of Transport Matters and Sustainable Australia.
While the Greens have potential to win more seats under the M5 system, they would have a more stable and safe number of seats under M40. They would likely hold 3 safe seats, a 4th mostly safe seat with the potential for a fifth in a good year.
You can see the difference for the Greens by comparing New South Wales - where they win 2-3 seats at every election - to Victoria or Western Australia where numbers have varied widely. You can also see it in the Senate, where the Greens are slightly over-represented now but a relatively small drop in the vote sustained over two elections could see them lose many seats.
This stability allows those small parties to build strength over time, as the Shooters, Fishers and Farmers have managed in New South Wales. In contrast, the rate of re-election for micro parties elected under GVTs is very poor.
Now New South Wales and South Australia - the only two at-large systems to date - elect half their upper house at each election for a four year term. I don't think that it's is possible to switch from a system where the entire house is elected at once to a half-chamber election system with such a long term.
Which is how Western Australia has now ended up with an M37 Legislative Council to be elected for the first time in 2025.
This brings me to my last point - once you are electing such a large number of members in a single constituency, a lot of the benefits of the Single Transferable Vote system get lost, and the difficulties in counting and voting become more pronounced.
If you're electing one, three, five or seven members, preferences are very valuable. There's a large gap between quotas, so if you simply elected candidates without preferences there would be a significant share of the vote unrepresented in any way. But as magnitude increases, that becomes less important.
On the other hand, as ballot papers get bigger, it becomes necessary to number more boxes for your vote to have its full value. Parties run more candidates. Even if you just number boxes above the line, generally more parties will run if the effective threshold for victory is lower. Perhaps no more than are running in Victoria now, but probably more than you would get in the medium term if Victoria abolished group voting tickets without any improvement in proportionality.
You have a number of solutions to the problem of ensuring preferences flow, and all of them have downsides. You can require preferences, imposing a higher burden on voters and punishing those who don't meet that burden with informality (or avoiding that burden with group voting tickets). Or you can not require preferences, and accept that some votes will exhaust. In either scenario, you are accepting that not all votes have the same power. Voters who are informed and educated enough to fill out the full ballot effectively avoid any exhaustion. Those who don't either cast an informal vote or have their vote exhaust.
There is also a higher administrative burden with counting ballots. It takes longer to count larger ballots - the person sorting the ballots has further to look to find a '1', and a person doing data entry spends longer on each ballot.
All of this can be solved by dispensing with using preferences, and instead using a party list system. These systems are common in Europe, and don't involve any preferences. Typical systems instead use a highest average method. Under the D'Hondt method, the first seat goes to the party with the highest vote, and their vote is then divided by two. This is repeated until all seats are filled, with the party's vote divided by the number of seats they have won, plus one. There is an alternative called Sainte-Laguë whereby the divisors go 1, 3, 5, 7 etc rather than 1, 2, 3, 4 etc, thus giving a small boost to smaller parties.
These systems allow for much simpler ballots and much simpler counting systems, but sacrifice the potential for preferences to come into play to decide the last seat. Sometimes parties might suffer from vote-splitting, but this effect is quite minor at the point where you're electing 21 or 40 members in the same district. In return, you achieve much greater simplicity both for voters and vote-counters, and ensure equal value of each vote without imposing a risk of informality on a voter.
In its simplest form, you could conduct a closed list PR election. The ballot could list the candidates for each party, or not, but it doesn't require boxes for each individual candidate. You just tick a single box and move on. We don't need data entry, you just divide up the votes between parties, and then apply the highest averages method to assign seats.
You could also use such a highest averages method but with a flexible or open list, which allows voters to cast a vote for an individual candidate. While this is currently possible, in practice voting below the line hardly ever changes the order of a party's candidates. In the absence of group voting tickets, below the line voting would become largely redundant. But you could set different thresholds that are more achievable to create the possibility of the popular vote influencing the party's list.
Some countries use a fully open system, where voters cast a vote for a particular candidate - the total vote share for the party decides that party's seat count, and those seats are then allocated to the highest-polling candidates within that party. Others use a flexible list, where seats are normally allocated on the list, but a candidate who passes a particular threshold of personal votes is promoted up the list.
These sorts of systems would allow you to cast a single vote, yet produce a very fair proportional result, with or without the ability to choose an individual candidate. Australians are very used to ranking on ballot papers, such that we often discount the high information burden this places on the voter, and the rates of failed ballots that would shock people in many other democracies.
Alternatively, you could keep a closed list system but still apply a preferential system which would largely work like above-the-line voting works in New South Wales now, but without the complexity of below the line voting. You could number parties, and a party would win a certain number of seats based on quotas with the final seats decided by distributions of preferences, but the ballot could be much simpler, with lists of candidates recorded in small print or simply available outside of the ballot paper.
The Single Transferable Vote is great when magnitudes are low, but when magnitudes are high I think Australians could look elsewhere for other models that would work better.
Moving to statewide election, potentially with a more appropriate electoral system, could resolve the issue of ensuring diversity in the Legislative Council without group voting tickets. I think it could win the support of some who would otherwise oppose reform.
The Animal Justice Party has been tweeting about its conditional support for group voting ticket abolition along with the introduction of "proportional representation". They haven't defined what this means (after all, PR includes the current M5 system) but I suspect it means moving to a more proportional system, like the statewide PR I have proposed.
Group voting has got to go but it has to be paired with proportional representation
— Animal Justice Party Victoria (@AJP_Victoria) November 13, 2022
Meanwhile the Coalition and One Nation have now also come onboard for reform, with Labor signalling that they will consider the issue after the election.
Such a reform (whether it involves a novel voting system, or simply abolishing the regions) would require constitutional change, I believe using a referendum. Some are reluctant to go for solutions that involve referendums, but I think the case can be made.
Voters know something is wrong now, and we can make a case that a different system would be fairer and more representative. So let's go beyond just talking about group voting tickets and talk more about what a better electoral system would look like.
I don’t think we need more proportionality than M5, and I don’t think we need a house of review. I think we can deliver better ‘governance’ through other methods, such as more use of citizens’ assemblies (newdemocracy.com.au). Why not attract support by proposing a reduction in the number of MPs? Consider a unicameral parliament with 95 MPs comprising 19 x M5 districts?
Or, if the consensus view is that M5 isn’t proportional enough, why not use M7, and thus retain the benefits of STV?
The diversity argument is not completely silly, but the solution is to pick up a number of citizen-provosts by random draw, not to rig the electoral system by using GVTs. Maybe 9 in the Victorian council. Give them relatively short staggered terms and a suspensory veto.
“When Machiavelli proposes a constitution for a revived Florentine Republic, he very subtly – indeed, almost surreptitiously – incorporates tribune-like offices, the provosts (proposti), into his plan. Machiavelli wrote the “Discursus on Florentine Affairs” (1519–20)30 in response to Giovanni de’Medici’s (Pope Leo’s) solicitations for advice on converting Florence from a de facto principality to a genuine republic, since the leading Medici, now Church prelates, will leave behind them no legitimate heirs to serve as princes of Florence.
McCormick. Machiavellian Democracy (p. 103). Cambridge University Press. Kindle Edition.”
Alan, you don’t need sortition. Just abolish the regions.
Jeremy, I’m not sure what benefits I see from low-magnitude STV in the upper house. Basically the only one is geographic representation, but the lower house does a better job at that anyway.
Once you’re unpicking the constitutional settlement I don’t see the point of just bumping up from M5 to M7 (even if you could work out the right arrangement that works for Victoria).
I would be strongly opposed to reducing MP numbers – smaller parliaments hinder diversity and are a pointless populist gesture that we should resist. And upper houses have their own distinct role in checking a government. The semi-parliamentary system we have in Australia allows us to have a separation of powers between an executive-dominated lower house and a more free upper house. You lose all that with unicameralism.
Citizens assemblies might have some value but it doesn’t fill the role that properly democratically elected parliaments play in holding governments accountable.
“I would be strongly opposed to reducing MP numbers – smaller parliaments hinder diversity and are a pointless populist gesture that we should resist.”
I couldn’t have put it better myself.
Agree Nicholas/Ben Raue, i support political diversity and feel reducing MPs is pointless especially as Victoria is fast-growing. Politics should be a spectrum not a binary choice. I think one of Bracks’ greatest Legacy was to reform the Victorian Upper House I wish that QLD re-established an Upper house as well
@Nimalan, agree with your wishes that QLD re-established an Upper House/Senate. The lack of an Upper House in QLD suppresses a greater potential of political diversity and most benefits the established major parties including the two-party system. Much like Victoria, its unlikely that any reform will be made in Queensland while there is a Labor majority in power. It will likely require a minority government with the crossbench putting the government of the day under some meaningful pressure until reform is actioned or put to the public.
The point about reducing the number of MPs is that it’s combined with unicameralism, where the assembly actually increases in size, e.g. from 88 to 95 (19 x M5).
The point about a unicameral solution is that it is based on M5 or M7 districts, where parties are usually required to compromise, i.e. form a coalition, to form government. The government is held to account by the fear of the coalition falling apart.
Instead of adding an upper house to QLD, it would be better to change their assembly to use PR-STV in 5 or 7 member districts.
Ben, the benefit of PR-STV is that representatives are directly elected.
@ SEQ Observer, agree that it is unlikely that while Labor has a majority government the issue will be looked into. A Minority government makes this more likely at least to get an agreement for a referendum as part of a confidence and supply agreement with the crossbench. Interestingly, in the 2016 referendum on Four Year terms the absence of an Upper House was cited by the No campaign. I would like Greens, KAP, ONP and some independents to sign a joint declaration that if ever there was a Minority government an agreement for a referendum would a prerequisite for any confidence and supply agreement.
Good afternoon Ben,
It would be beneficial for voters to vote below the line, however, it’s critically important for voters and candidates to have an easier way to connect if you want voters to number more of the boxes. There is no point having undecided voters tick more boxes, if they don’t know WHO they are voting for.
Vote Easy is the perfect vehicle for that. It operates nationally, provides a level playing field, affordable and delivers candidates campaign directly to voters, no third-party interference.
Whatever system the commission uses, Vote Easy will help make it easier for the community to connect and make informed decisions.
Regards
Craig Blakley
Jeremy, consider the NSW Legislative Council. Those MPs aren’t “directly elected” in any meaningful way that is different to MPs elected under list PR. I don’t think there’s much difference between that and the Senate or the Victorian upper house. Whereas list PR could potentially involve the ability to alter the party list.
Ben, how would a system like MMP used in NZ work for some of the state legislatures? Queensland as a unicameral legislature could be converted to MMP with 150 seats or so. 75 would be elected from single member districts as currently and then the other 75 elected using a list system. It would be a bit more difficult since it may not work well with preferences involved.
The split could also be 90 from single member districts and then 60 from the party list
Eh I’m not a huge fan of MMP. In particular I don’t like the all-or-nothing cut-off at 5%, whereas there is more of a gradual natural threshold under STV. You could theoretically make the local districts preferential but not sure there’s a point since they don’t have any influence over who forms government which is the other thing I don’t like – they have a whole bunch of elections which are rendered pointless by the list.
There’s absolutely no necessity for a threshold percentage in MMP and under high fragmentation it causes obvious problems. It’s explicitly there to weed out small parties – and as we’ve seen it has various issues (see ACT and NZF in NZ generally, and the FDP in Germany being at threshold level). There’s an argument about keeping extremists out with the threshold, but if you find that persuasive then why bother with more than two parties in the first place?
Making (keeping!) local districts preferential, or at least not FPTP, under MMP is worthwhile per se, because under a more proportional system you’d expect the majors to be less dominant and have more three and even four-cornered contests as a result. It might also be worth shifting district to a Condorcet method.
If a threshold is insisted on, I think it would be worth pairing that with the list vote being optional preferential (run-off until everyone left is over the threshold). Supporting proportional representation is equivalent to supporting (almost) every voter being able to say “my vote helped elect X”.
More spicily, with preferences we can do single-ballot MMP.
—
Back to Victoria, out of the open-list PR systems the Dutch one is pretty great.
If I may, a couple of comments on Jeremy Lawrence’s comments:
Coalitions – For every country with a stable system of PR government, there is another that is completely chaotic. This of course makes sense, once you need coalitions to govern you increase exponentially the power of the smaller partner. If you have to compromise, and the issue is one where the 2 parties can’t agree, then you have a recipe for chaos. Even when there is a longstanding coalition, as with the Libs/Nats here, you can see this start to break down due to the same issues (Teals anyone?).
Citizens Assemblies – Perhaps the worst idea ever (only direct democracy comes close). These are never genuine ‘citizens’ assemblies, either they are drawn entirely from the ‘expert’ class, or they operate like Grand Juries in the US, in theory a disparate group of citizens but in practice entirely guided by a set of ‘experts’ to reach a pre determined conclusion. Even when working exactly as they are described, you have a bunch of people who have little interest in governing doing exactly that. Maybe if you got a bunch of people who were interested in governing, then had a vote to determine which of those went to some kind of assembly…
Personally, I think we have it right in this country – AV for single member lower house constituencies and PR for the house of review – but I do think a single statewide vote would be better. For example, if I live in a suburban area I may want to vote for a regional based party but may have no opportunity to do so if that party doesn’t run in my region (or rural and want to vote for an urban party).
My wish list would be we first choose the number of MPs we need. Cube root law for Victoria’s population gets up to 188 total (no need to be scared – this idea that politicians are a drain on society and need to be culled is ridiculous). We can probably trim that back a bit as it is a sub-national parliament. Maybe let’s say work out the cube root and then take 80% of that. Gets us to 150. Then proportion of of lower house to upper house members. Let’s say 75% / 25% split. That could get us to a 38 seat upper house and a 112 seat lower house. A 112 seat lower house would give 16 seats elected by STV M7 (NO above the line – Tasmanian Hare-Clarke system) which would be reasonably diverse while ensuring there aren’t too many parties, rural seats could be fairly cohesive (by that I mean there wouldn’t be seats stretching from the Wimmera to Gippsland etc), and government (provided our politicians act like adults) could be formed without too much difficulty. Then per Ben’s suggestion, elect the 38 seat upper house by St Lagüe or D’Hondt from a statewide electorate for exactly the reasons Ben states. That would give minor parties with a clear agenda that is supported by a significant section of the community a chance to grow without being subject to a lottery every 4 years. Politicians from those parties that prove themselves and build a profile would have a real chance at moving to the lower house. A more dynamic and responsive electoral system. I’d also make it harder for what I call “short term coalitions of the disgruntled or rent seeking” to get on the ballot in the first place. Instead of a small number of members having to be actual paid members of a party I’d raise the bar much higher but instead of members I’d make a process of giving a formal expression of support for a new party to be on the ballot. That might mean logging into a electoral commission website and giving a few details but would be just like signing a petition to say that this party should be on the ballot. People do online petitions every day. A new party can campaign however they want to drum up support for their right to be on the ballot but they need to get the signatures – digital or otherwise. The electoral commission would be tasked with checking the signatures are valid and not gamed. I don’t believe we need much more than around 10 legitimate parties to cover the spectrum of opinion in our society and smaller ballots would make voting easier for all concerned.
Comments are closed.