Senate voting reform looks set to go ahead


According to a report in this morning’s Guardian, the Coalition, the Greens and Nick Xenophon appear close to an agreement on Senate voting reform. There’s also an accompanying media release from Lee Rhiannon.

The proposed plan, put to the government by the Greens, would abolish group voting tickets (the mechanism whereby preferences flow according to pre-lodged party preference decisions when voters vote ‘1’ above the line for a party), and would allow voters to number their own preferences for parties above the line or for candidates below the line.

This proposal is very similar to that proposed unanimously by the Joint Standing Committee on Electoral Matters (JSCEM), with a few exceptions:

  • Voters would be expected to number at least six boxes above the line or twelve boxes below the line. As I understand it, there would be a savings provision so votes that don’t number enough boxes would still count.
  • The Greens don’t propose any changes to party registration rules, whereas JSCEM had proposed raising the threshold for party registration from 500 members to 1500 members.

While the article is written as if agreement has been reached, it’s unclear whether the Coalition has agreed to the Greens’ demand that party registration be taken out of the package. It’s also unclear where Labor stands on the proposal – while some senior Labor senators are opposed, there are others in the party who support Senate reform.

While some have been strongly opposed to any changes to group voting tickets, a lot of anxiety around Senate reform has focused on the likelihood that most voters would just vote ‘1’ and this would lead to a lot of votes exhausting, and thus the final seats being decided on substantially less votes than a full quota, and thus distort the result.

This proposal would partly mitigate against this, although a voter could easily vote for six tiny parties and still have their vote exhaust, or even just vote ‘1’.

I suspect that the legislation will come with a provision to also require parties registering how-to-votes to include six preferences. Federal legislation has a history of allowing a particular type of vote to be formal, but requiring the ballot paper instructions, and any campaigning, to advocate for a higher standard of formality. This kind of clause is controversial, though, because it can lead to criminalising advocating for what is effectively a legal vote.

Albert Langer’s case in the late 1990s was due to him advocating for a type of voting which at the time was legal, but his advocacy was illegal. The Parliament responded to this case by eliminating the voting option, and thus bringing the voting rules into line with what was legal to advocate.

So what does this mean for preferences? Firstly, they’ll become much less important. Even if there’s a requirement that parties number six boxes, parties will only be able to direct preferences if they can put a how-to-vote card in the voters’ hand, and still rely on the voters following that advice and filling it in themselves. This means most microparties’ preference recommendations will become worthless, and major parties will not be able to have as much flexibility as they do now.

For example, if even if Labor and the Greens didn’t preference each other, it’d be a lot easier for their voters to override this advice and give preferences.

Whether or not there is a requirement that parties give more than one preference on their how-to-vote, there will still be an incentive to encourage your voters to preference and to do deals with parties which have something to offer, but those preference deals will have much weaker effects and will be much easier to understand.

There will still be a lot of collective muscle memory after over 30 years of voters just voting ‘1’ above the line, and you’d expect plenty of voters to still vote ‘1’, even if parties recommend otherwise.

If there is no rule against advocating ‘Just Vote 1’, you could imagine the LNP adopting a similar strategy to recent NSW and Queensland elections to try to reduce preference flows between Labor and the Greens, although this would be risky in an election where voters are required to number every box on their House of Representatives ballot.

Preferences can still matter in close contests, and you’d expect that if parties are required to number six boxes on their how-to-vote that we would still see more preferences flow, but they would definitely be less decisive than under the current rules – the candidate leading after early rounds of counting would be the clear favourite to win.

(When there is more time, I might do some analysis of what preference flows look like in NSW elections for parties that have chosen to issue preferences – I think a lot of their voters still exhausted.)

It’s also interesting that the Greens have rejected any changes to party registration rules. Before the last election, in an attempt to address the surge in candidates running in federal elections without making any fundamental changes to the system which encouraged the surge, the major parties voted through big increases in nomination fees. This failed as a strategy, with the number of parties doubling and the number of candidates increasing significantly.

The major parties have still supported using increased membership requirements as a parallel strategy to reduce the size of the ballot, but I’ve personally always thought it’s an attempt to address the symptom without dealing with the disease.

There are a number of parties which rely entirely on chance preference arrangements, and would likely fold if the prospect of getting elected from preference harvesting goes away. Even for those parties which remain, the incentive to run in every single state in order to swap preferences will go away. If your preferences only matter when you hand out how-to-votes, it makes sense to only focus on states where you have members.

If this reform is passed, it has serious implications for the prospect of a double dissolution.

It’s very difficult to hold a double dissolution under the current rules – ballots would get even bigger, and preference deals would allow even more candidates to be elected on minuscule votes. It would also add pressure on the AEC, since there would be more close contests which could produce recounts or revotes.

A double dissolution would become easier if Senate reform is passed, but it would also make it more necessary. Apart from John Madigan, whose term expires in 2017, the rest of the crossbench will remain in place until mid-2020 – and legislating Senate reform will make it even more difficult for the Turnbull government to work with these senators. A double dissolution would take that problem away.

Timing is a problem for a double dissolution. The AEC will need some time to modify its software and prepare different counting processes to deal with a lot more ballots which will have more than one number written on them.

If a double dissolution is held before the end of June, the next half-Senate election would be due by the middle of 2018 – effectively cutting a whole year off the next term of Parliament.

A double dissolution could be held on the 2nd, 9th or 16th of July, but these dates would require a longer-than-usual campaign, since the double dissolution must be called by May 11. They also clash with the mid-year school holidays, which is why we’ve had so few July elections.

None of these options are ideal, but if the Turnbull government really has burnt its bridges with this Senate, they may judge a July double dissolution to be necessary.

Liked it? Take a second to support the Tally Room on Patreon!
Become a patron at Patreon!


  1. If this change goes through, it may well have a flow-on effect in NSW Upper House elections as people learn that putting more than one number above the line is valid and worthwhile.

  2. From a practical point of view, a double dissolution almost immediately following a major change to the Senate voting system would be a bit like a bungee jump with an unknown length of rope. The last time there were big changes to the Senate voting system, they were enacted in late 1983, came into force on 21 February 1984, but didn’t have to be used at an election until the 1 December 1984 poll. And of course the senators didn’t actually have to take their seats until 1 July 1985. (I remember this well, as I was on the AEC team that wrote the procedures for the new system and delivered nationwide training, including for political parties.)

    As to whether the AEC, given amendments passed in, say, March this year, could have them all computerised, tested, certified and ready to be used for a July election following a double dissolution, well, frankly, nobody knows. Even if they could, how much data entry might be needed will depend on the specific ballot paper instructions and the numbers of groups and candidates in each state – and maybe also on what parties the can and do put on their how-to-vote cards. And if the great bulk of ballots have to be data entered – which hasn’t happened before – it’s anyone’s guess whether there are even enough skilled data entry operators in the labour market these days to enable that to be done in any sort of reasonable time frame.

    I could easily foresee a scenario where the Senate counts took two months to complete, as happened at the notorious first ACT election in the late 80s. And that would create big problems, not least because we could be without a Parliament for close to four months. I’m not sure of the implications for the budget and supply, but there would be a problem for Senate ministers, since under section 64 of the Constitution they can’t keep their positions for longer than three months without being a member of either the Senate or the House.

    And if the counts had to be done manually, it would be even worse, for reasons I’ve discussed in a comment on Antony Green’s blog.

  3. I don’t mind these changes although I do have some reservations.
    Firstly, I would prefer a NSW style where people can choose to number as many boxes above the line as they desire.
    Secondly, voters could number a minimum of 6 boxes below the line in a regular half senate election or 12 in a DD election, rather then currently fill out every box.
    Furthermore under these proposals, especially when voting above the line, current formal votes would be actually informal, and with reference to Antony Green’s contribution on his blog, this could be a repeat of the 1983 election in which their was an informal vote of 10% for the senate, so making a savings provision by allowing optional preferential above the line would be smart.
    Finally by keeping the proposal of numbering 1 to 6 above the line could have a consequential affect on lower house informality rates, as voters may confuse to 1 to 6 voting pattern with the lower house rules of full preferential voting.

  4. Double dissolution speculation is the most over used electoral expression of the last ten years.

    I think with the last polls we can (hopefully) agree that there will not be one this time around. If i’m wrong feel free to remind me about it constantly when i make predictions. Now stop the speculating.

  5. Andrew, I would normally agree with you. Double dissolutions are often discussed as shorthand for ‘early election’ – however I think if there’s a DD it won’t be to go early – if anything the election could be just a month earlier than a standard election – but in order to clear out the Senate, which the government is clearly incapable of working with.

    When you think about the DD as a method to flush out the Senate, not as an early election trick, it makes a lot of sense.

    I don’t see how the Coalition can live with this crossbench for the next four years, particularly if they pass Senate reform.

  6. I’ve been out of the country for some time so Ill admit a lack of time and focus on Australian politics. I don’t see a completely new senate being much better for the coalition though. The Liberals only lost one spot last time to Ricky Muir. The rest came at the expense of Labor. So they have their ‘ideal’ Senate bar one spot.

    In regards to a normal half election:

    Perhaps they believe they are entitled to the NSW spot of the Liberal Democrats, but its hard to imagine a repeat delivering a whopping four (out of six) spots to the Liberals in NSW particularly with an existing Liberal State Government.
    If I was doing my early senate speculation I’d say they are likely to do about as well as they did 2013 or 2010 under the current system. With the caveat of Victoria being better for them, but maybe an unlikely outcome in South Australia of a bit worse.

    A half election with the proposed new rules:

    I’m not sure what will happen in practice with preferences, but I expect that the Liberals will be able to normalize getting their three spots in Victoria and any bonuses for them coming from the more left parties making a mistake between them. Once again they might also do worse in South Australia. but getting one spot really would be a complete stuff up result for them.

    A full senate election under the new rules.

    I’m sure that there are others with far better predictions than me, but I suppose that they might hope for more than the 50% of spots in WA and actually getting 50% in South Australia. Id say they would have problems in the territories and Tasmania with more chance of Greens and Labor swapping preferences and actually getting over the 50%. I don’t see a really positive risk return outcome here.

    Am I missing anything?

  7. Obviously voters and parties would change their votes and preferences based on a double dissolution and a new voting system, but first preference quotas based on recent half senate elections if they carried through to a full senate election:

    One quota for Labor and Liberal and close enough to one for the Greens, leaving the last spot to be fought out between Labor and Liberal or a very unlikely preference accumulation.

    Four Lib, four Labor, and just over one for the Greens and Liberal Democrats. Taking the Liberal Democrats result as an anomaly and 2010 as a better example would put the Liberals on about five and Labor and the Greens a bit over again and closer to a second spot each on preference flows.

    Would have been two for Country Liberal, another for Labor with the last coming down to preferences.

    Once again looking more at 2010 its over five for Liberals, practically four for Labor and in a more 2010ish result than 2013 another for the Greens (who would be very highly placed to take one spot on preferences if it came down to that anyway)

    Three for Liberal, just over three for Xenophon (instead of one!), three for Labor and almost one for the Greens.

    Depends very much on a 2010 or a 2013 result. But potentially Four or Five for either Labor or the Liberal and one or two for the Greens.

    Once again looking at the winner getting five the second place holder four and the Greens practically getting one full quo up to almost two.

    Straight five for Liberals (with plenty left over if they can manage preferences with Nationals) almost four for Labor and between one and a half to close to two for the Greens.

    I’d say just on that its a negative for the Liberals in ACT as they risk one spot with zero potential of picking up another. NSW they’d still have to pick up their sixth spot on preferences or attrition and there’s the chance of the Greens improving to two either at the expense of or addition to Labor’s four. NT not much of a difference. QLD they’d probably get their six and could hope something went wrong for everyone else. SA Xenophon has gone to three from one which I wouldn’t call a win for the Liberals. And that’s just on quota I wouldn’t claim to know who to predict for the last spot. I suppose on preferences and attrition they could hope to improve to five in Tasmania ( I cant see them doing worse than four) and they probably see senate reform as a fix for VIC to get their six spots which is about as good as they can hope for in WA.

    So more or less risking one quota in ACT and giving a current cross bencher two more quota spots to deal in SA. In QLD and NSW the one Green spot each is already going to ballot so a double dissolution would just make it easier to keep the status quo (and those spots would just go to Labor anyway)

    Seems like a lot of work to get rid of Ricky Muir and PUP

  8. ACT is clearly one seat for the ALP, Liberals continue to get just over a quota or close enough to it to get there . The proposed changes to the voting system I suspect will have less impact in the ACT than anywhere else. It would have prevent the sort of payback on ACT Greens exercised by the Animal Justice Party in the last election whose ticket preferenced the Liberals above the Greens over kangaroo culling in the ACT if you can believe that.

  9. The removal of GVT will be all downside for the Greens in the ACT Senate election. Against the Liberals for the second spot, the Greens got the preferences of every ticket bar Rise Up Australia and Animal Justice at the last election. Exhaustion and increased leakage under the new system will make a Liberal victory assured.

    (I still support the removal of GVT).

  10. I see, so ACT NT is two seats for the term or the other two seats stay sitting?

    That means potentially another Liberal spot in TAS and VIC along with a less likely increase in WA and SA. But they give the Greens more chance of improving from one to two in QLD and NSW and a better guarantee of at least going even there and very likely hand the cross bench SA seats over to Xenophon. The potentially worse situation for the Greens in WA, SA and TAS is very likely mitigated as it would probably be a fight between them and Labor for the last spot rather than it going to the Liberals.

    And that’s just the parties polling well on quota, there’s every chance of another ‘new’ fourth party polling 4% plus which would likely be irrelevant under the new model at a normal election, but would still have a very good chance with the lower quota at full senate election.

  11. Perhaps they have the same opinion as the experts mentioned in this article.

    NSW Coalition 7, ALP 4, Greens 1
    Vic Coalition 6, ALP 4, Greens 2
    Qld Coalition 7, ALP 4, Greens 1
    WA Coalition 7, ALP 4, Greens 1
    Tas Coalition 6, ALP 4, Greens 2
    SA Coalition 5, ALP 3, Greens 1, Xenophon 3
    ACT Coalition 1, ALP 1
    NT Coalition 1, ALP 1
    Total: Coalition 40, ALP 25, 8, Xenophon 3

    I think that predicting seven Liberal in NSW and QLD is utter nonsense. On NSW 2010 results they get five on quota with their sixth spot behind the second Green spot. A 2010 type result for Labor in NSW would also potentially put them in a spot to compete with the Greens for that last spot. It would be hard to see where the Liberal preferences would come from to get them ahead of either for a seventh.

    Even on QLD 2013 results the LNP end up on five and a bit quota, Labor close enough to four and the Greens close enough to one. So giving Liberal six, labor almost four and Greens almost one that leaves the entire quota and surplus sitting with Palmer at the moment. Either some of that primary vote goes back to Labor or the Greens taking them over quota and hence competing for that last spot with another minor party or it goes to that minor party and they will be likely to win the last spot. Where do the experts think that the seventh LNP senator is going to get their primary votes from? An unprecedented increase from 41% to 47%+ primary vote?

    SA and WA do look worse for the Greens, but at least in the former case the preferences between Greens and Labor might make it another competition between them for the last spot. Current preferencing between Liberal and National would help them anyway regardless of any changes in WA.

  12. That article is garbage. If the Coalition get 46% first preference vote in a state, they can win 6 seats and no more. They would elect 5 Senators, reach the sixth quota, and then be excluded. It might be possible to win 7 in WA in a good year, but that’s about it.

  13. I was about to say the same. I don’t know the other guy, but Peter Breen is clearly conflicted by his involvement in microparties – and the article identifies the other guy as likewise having a role in negotiating those preference deals which will become redundant.

    Apart from Antony’s point about the absurdity of the Coalition getting 7 senators in three states, the remarkable confidence with which they predict zero other minor parties from getting elected – with a minimum of six preferences and some parties getting 3-4% and a 7.7% quota it certainly is possible.

    Also their great confidence that the Greens can’t get two elected in WA is interesting considering how well the Greens polled there in 2014. Sure, it’s possible but they demonstrate remarkable confidence.

    The latest BludgerTrack quarterly has the Liberal Party on 46.5% in WA – so basically nothing left over after six quotas, even if you assume there’s no drop in Liberal vote in the Senate, which there always is.

    The ACT and NT only have 2 senators each, full stop. Those senators serve three year terms, so a DD would function exactly the same as a regular election for the territories.

  14. A full senate election, new preference system or not, is actually the best case scenario for minor or micro parties. Just reducing the quota to around 7.7 puts a number of them at just under or just over half a quota rather than a dismal 25% or less. Easy to possibly imagine a micro party or two (particularly if they get some version of a donkey vote) battling it out for a last spot either against each other or Labor or the Greens.

  15. The issue of Senate term dates is actually more complicated in the case of a DD around the middle of the year because that is around when section 13 of the Constitution starts the Senate terms of each state/each Senator (it is not entirely clear whether the terms are individual or by state but this matters little to not at all since proportional representation and then the Casual Vacancies amendment have made it nigh on impossible to have the Senators from a state elected on different days), based on the “day of his election” and it has never been ruled whether this means polling day, the day the polls are declared or some other less likely day such as the return of the writs.

    So, if the day of the declaration of the polls is “the day of his election”, then an election on the 25th or probably 18th of June would result in a longer term and an election on the 11th would result in a short term, unless the count dragged on.

    Also, regardless of whether it is polling day or the day of the declaration of the polls that constitutes “the day of his election” when there is a contested election, there is a strong case that, in the extremely unlikely event of, a state having exactly 12 candidates standing in a DD, “the day of their election” is nomination day and thus short terms are certain in that state.

  16. If the Senate count was delayed, it would take a shorter time in some states and territories than others, allowing the Senate to commence before all the states have elected their senators. Under the Senate (Quorum) Act 1991, only a quarter of Senators are required and that only needs 2 states. The territories would almost certainly finish the election first, as they have the smallest and simplest counts and then this would likely be followed by the smaller states such as Tasmania and South Australia.

    A delay in the Senate election would not effect the ability of the House of Reps to sit and thus allow, if the parliament was not unworkably hung, there to be a government with the confidence of the House.

  17. I think that the territories should get an extra Senator each. This would eject some competition into their Senate races, particularly that of the NT. The ACT would be a bit more stable with 1, ALP, 1 Lib and 1 Green.

  18. Askey was the ‘numbers man’ behind the HEMP Party and a key architect of the preference dealing with Druery that saw parties like HEMP and Wikileaks accidentally help elect a Liberal over the Greens in NSW in 2013.

    Askey (Registered Officer) and Breen (Secretary)’s application to register the Renewable Energy Party has hit the AEC website today

  19. @Tom, I’ve thought for a while that we should increase the number of Senators per state to 14, which would have the benefit of meaning the median seat in each state would be decided by one side getting to 50% instead of the current stasis where usually the left and right are stuck on 3, and would add roughly 25 seats to the House of Representatives.

    If you did that it would also make sense to add one seat each to the Territories, although that wouldn’t change the House.

  20. Totally agree with the increases as Ben suggests. The Nats would presumably support it (as they did in 1983) because it would help offset the loss of regional seats due to population shifts, so I don’t see it being as politically impossible to implement as some may think.

  21. Breen/Askey are serial offenders with this sort of rubbish “analysis” and journalists should not pass them off as relevant authority figures.

  22. @NickC,

    I don’t believe that expanding the Parliament would be “politically impossible” at all. On the contrary, every party would love it.

    Major parties would love it, who doesn’t love the prospect of winning more seats?
    Nationals would love it, as you say, because it preserves rural seats.
    Greens and Independents would love it, because smaller seats allow greater concentration of their vote, making it easier to win.

    And electing 7 Senators per election would be reasonably well supported across the board too, because it potentially removes the 3-3 deadlock, while also giving minor parties a better chance to win with the lower quota.

Comments are closed.