Senate preferences – how might they flow? Part 2


Following yesterday’s post analysing the NSW, VIC and QLD Senate counts, here’s my analysis of Western Australia, South Australia and an update of the Tasmanian race.

Western Australia

Full quotas: 5 Liberal, 3 Labor, 1 Greens

Parties in the race:

  • 0.7027 – Labor
  • 0.5025 – One Nation
  • 0.3378 – Greens
  • 0.3311 – Nationals
  • 0.2757 – Nick Xenophon Team

Votes to be distributed:

  • 1.1226 – right
  • 0.4547 – left
  • 0.1792 – random
  • 0.0934 – centre

The fourth Labor candidate (Louise Pratt) should win the tenth seat, leaving One Nation, the Greens and the Nationals competing for the last two seats, with the Nick Xenophon Team as an outside chance.

There are substantially more right-wing preferences than there are left-wing minor party preferences, including 0.23 quotas for the Shooters, 0.23 quotas for the Australian Christians, along with votes for the Australian Liberty Alliance, the Christian Democratic Party, the Liberal Democrats and the DLP.

These votes may well favour one of the two right-wing contenders strongly, and the other may find themselves stranded behind the Greens. For those following preference recommendations, Shooters preferences favour One Nation, while how-to-votes for the Australian Liberty Alliance and the Australian Christians don’t recommend a preference for either the Nationals or One Nation,

On the left, 0.22 quotas went to the HEMP/Sex Party coalition, with 0.11 quotas for Animal Justice and another 0.12 quotas for four small parties. Both HEMP/Sex and Animal Justice how-to-votes favour the Greens.

The Nick Xenophon Team could also be a contender if they gain a strong number of preferences, but would otherwise have a lot of preferences to distribute.

South Australia

Full quotas: 4 Liberal, 3 Labor, 2 Nick Xenophon Team

Parties in the race:

  • 0.8340 – Nick Xenophon Team
  • 0.7153 – Greens
  • 0.6001 – Labor
  • 0.3866 – One Nation
  • 0.3601 – Family First
  • 0.2686 – Liberal

Votes to be distributed:

  • 0.4003 – left
  • 0.2606 – right
  • 0.0891 – centre
  • 0.0851 – random

The third Nick Xenophon Team candidate, the fourth Labor candidate and the first Greens candidate all have a clear lead to win the last three seats.

Family First and One Nation are next in line. One Nation is unlikely to benefit from preferences, but Family First is second on the Liberal how-to-vote, which you’d expect to have been followed by most Liberal voters.

You’d expect less discipline from the Shooters (0.09 quotas), Liberal Democrats (0.08) and, if Family First get ahead of them, One Nation (0.39), but enough of them will favour Family First to make them a contender.

On the left, preferences from the HEMP/Sex Party ticket and Animal Justice should help both Labor and Greens, but would mostly flow to the Greens, likely putting them in a clear winning position, alongside Nick Xenophon Team.

Labor’s fourth candidate sits on 0.6 quotas, and is probably the favourite for the last seat but could be at real danger from Family First.


I explained the unusual situation in Tasmania last Thursday, and not a lot has changed since then.

We now have complete data from 54 booths in Tasmania – which make up about 14.5% of the total ordinary vote across Tasmania counted so far.

Lisa Singh now has 20% of the Labor vote, and Richard Colbeck has 13% of the Liberal vote. Colbeck is doing best in Braddon, which is unrepresented, while Singh is doing best in the overrepresented southern parts of the state. When you adjust for electorates, Singh has 16% of the Labor vote and Colbeck has 14% of the Liberal vote.

I’m assuming that other votes mostly stay with the primary vote party for this analysis.

Full quotas: 3 Labor, 3 Liberal, 1 Greens, 1 Jacqui Lambie

Candidates in the race:

  • Lisa Singh (ALP) – 0.7239
  • Catryna Bilyk (ALP) – 0.6489
  • David Bushby (LIB) – 0.6016
  • Richard Colbeck (LIB) – 0.5899
  • Nick McKim (GRN) – 0.4119
  • Kate McCulloch (ON) – 0.3371
  • Peter Madden (FF) – 0.2533

Votes to be distributed:

  • 0.4975 – right
  • 0.4363 – left
  • 0.3426 – centre
  • 0.1565 – random

The two Labor candidates are in lead, followed by the two Liberal candidates, with the Greens, One Nation and Family First trailing behind.

Amongst minor candidates, there’s about half a quota of right-wing votes and almost as much in left-wing votes. Family First and One Nation are likely too far behind to overtake a Liberal. Most of their preferences would be above the line, which means they can only favour Bushby over Colbeck, until Bushby reaches the quota. This could mean that Colbeck is starved of preferences.

Having said that, only about one third of the below-the-line vote was for Singh and Colbeck, so there could be plenty of other preferences for minor candidates including the sixth Liberal.

The left minor party vote includes 0.17 quota for the Sex Party/HEMP ticket and another 0.09 quota for Animal Justice, and you would expect these to favour McKim more than the other tickets, but he would need most of these votes to overtake Colbeck, assuming he gets few preferences of his own.

Bushby, Singh and Bilyk should win seats, and Colbeck looks more likely to win than McKim.

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


  1. Wow that’s brutal for McKim. If half the Greens had gone below the line for him first both Green candidates would easily coast over the line with 0.7 quotas.

    If exhaustion does prove to be routinely high with the new system strategic voting will be imperative for parties to maximize their chances at seats.

    I mean consider the Liberals in WA in a half senate election. If the Liberals primary ~2.69 quotas, but spread it 4 ways that’s 0.67 quotas each! The Nats and every other right wing party would be crowded out and it’s quite possible the Libs win 4/6 seats, with just over 1/3 of primary vote!

  2. Yes, under the new system in this DD it might’ve actually made more sense for a party with a vote around the level of the Greens to have actually run two separate tickets in most states and tried to split their vote evenly between them. They may have won extra seats!

    If the Labor vote in Tasmania had been all ATL, they’d clearly have much less chance of 5 seats than what they have because of their vote being split.

    FWW there is a way of avoiding this potential for strategic voting to distort results – use a counting method like Schulze STV

    It’s worth serious consideration because strategic voting campaigns can be risky and unpredictable for the parties attempting them, and pretty bad for any parties on the losing side when they come off, but a system like Schulze STV counteracts strategic voting.

  3. The vote probably would not have been evenly distributed between the Green groups if 2 tickets had been won.

    Multiple tickets dramatically increase preference leakage by allowing in-party ATL leakage.

    Multiple tickets create confusion for voters and could be seen as tricky, lowering the vote.

    It would have done nothing good and potentially been harmful for the Greens in Queensland, NSW and SA (to be fair, it can very easily be done in stronger states but not weaker states).

    It would not appeal to those who want to know which of a parties candidates will get elected (which in case of political parties often includes the party itself).

    And most crucially, it could only help the Greens in the short Senate term and could harm them in the following half-Senate term because of the different methods for choosing who get which terms after the DD.

    There has apparently been talk that the first elected method, which has always been used previously (even though there was a new method available for after the 1987 DD), will be used. The Greens running split tickets would mean that neither Green would have a full quota at the start and thus, under the first elected method, would both get short terms (if both elected).

  4. Does your estimation of Singh’s position take into account Labor voters who voted below the line and gave Singh a vote somewhere between two and five? I think a number of Labor voters will have skipped Short or perhaps skipped Polley (with her anti-marriage-equality views) and given Singh their no 5 vote. I don’t these votes will become apparent until preferences are distributed.

  5. Only if they preference Singh ahead of Bilyk will BTL ALP voters, who voted for 1 for any of the top 4 ALP candidates help Lisa Singh against Bilyk.

    If they preferenced Bilyk ahead of Singh the part of their vote not remaining with top 4 ALP candidates will go to Bilyk and stay with Bilyk unless Bilyk`s votes are distributed, regardless of whether or not they skipped Short and/or Polley.

    In fact skipping Bilyk and/or Polley actually helps Bilyk slightly because a greater proportion of those votes goes to Bilyk because none of it is left with the skipped candidates, causing a slightly greater proportion of each of the ALP BTLs that do not go to Bilyk ahead of Singh to stay with Short and/or Polley. (I have not however accounted for the use of Unweighted Inclusive Gregory instead of Weighted Inculsive Gregory).

  6. I have been sampling Tasmanian votes via scrutineering during the scanning process, but it is very time-intensive as most of the votes are irrelevant and even when a vote is relevant you have to be quick to get the relevant data before it is cleared and vanishes off the screen.

    The #1 Urquhart votes mostly flow down the ticket, or if they put Polley 6th (which a few do) they still usually have Bilyk ahead of Singh. There is a “leak” there but it isn’t massive; it seems to be smaller than the roughly 20% flow from Abetz BTLs straight to Colbeck.

    The BTLs that are #2 for candidates below Urquhart are a different matter – quite a lot of them go to Singh. The impact of flow from Polley and Brown to Singh is reduced by Inclusive Gregory, but this is not the case for Short.

  7. I wouldn’t be surprised of voters down in Dension, whom would have voted Wilkie in the lower house may have voted for Singh below the line whilst also preferencing McKim right behind her, I could imagine a lot of Labor and Green leaning voters doing as well. In the end it will be close between Colbeck and McKim, i’d think that Colbeck will come out on top. Although when it goes back to the usual half senate election, that will inherently favour the Greens, and if they lose this second seat they win it straight back next election.

  8. There is a lot of 1 Singh 2 McKim, and there are booths like Fern Tree (a traditional Green haven) where Singh got 18% and the Greens were down 11%. I knew a lot of people who were tossing up between the two. People found the novelty of saving a left-wing Labor senator from the hacks far more appealling than a rather routine Greens campaign.

    My current model ( suggests it’s possible both Colbeck and McKim could win at the expense of Bushby, mainly because of the level of 1. Abetz 2. Colbeck BTLs. The Liberals’ ATL performance on the right micro preferences is not that flashy in my sampling so far.

  9. I presume that Singh will get a short term and thus will be back at the mercy of the factions next time, with a much lower chance of beating the group candidate order.

    From what little I understand of Tasmanian Liberal factions, basically Abetz is from the pro-Abbott Right (which is dominant in Tasmania at the Commonwealth level) and Colbeck is from the more moderate pro Turnbull part and that is what caused his demotion (despite/because he is the only Tasmanian minister in the Turnbull cabinet), 1 Abetz, 2 Colbeck seems like an ideologically mixed vote.

    It is good that the Tasmanians have their partial choice of major party Senator back, at least in DDs, with the new system.

  10. Tom, 1 Abetz 2 Colbeck is a “vote for the people I’ve heard of” vote. Apparently there’s some number of BTL votes that seem to go for the most recognisable names completely across party lines.

  11. That does seem to be the likely answer. I had thought of the “I`ve heard of” vote but forgot to mention it.

  12. Not possible, as far as I see it. Short has got basically no BTL votes so only get votes if Bilyk has gone over quota, and there’s not enough votes for her to do that if Singh comes close to what she’s currently polling.

  13. The BTL votes for Singh and Colbeck seem pretty set in now, moving around by tenths of a percent when rejigged each time but no more. On that basis Short has absolutely no chance – Labor minus Singh can’t get four quotas so Short will be eliminated very quickly (he’s on less than half a percent).

    I don’t have anything in particular against Bilyk but she is probably the most invisible Tasmanian Senator in decades, and that’s saying something.

  14. It will be a while before the first preferences for all BTL candidates are known. But Cash is currently on 223 and should finish well into four figures, maybe 1500 or so.

  15. Just pushed the button and according to KB, the result is 5 Lab, 4 Libs, 2 Greens, 1 Lambie. Good result for left with 7. Disaster for Libs with only 4 and Lambie opposed to a number of Govt policies. So almost an 8-4 result against the Libs.

    Wonder how Abetz will spin this?

  16. For all that intrigue in Tassie the state still returns the same amount of senators doesn’t it? Good result for Singh, don’t think that heavy a BTL vote could happen in any other state.

  17. The last seat in Tasmania didn’t fill the quota, as expected in almost every state. On the last count McKim won with 21,247 votes, narrowly edging ahead of McCulloch who had 21,106 votes.
    The Greens were lucky here, they should have achieved 2 quotas in their own right, I dare say that the Singh factor really hurt them.

  18. Funnily enough it was the distribution of votes held by Colbeck when he was eliminated that pushed Singh across the line.

  19. Suggestion seems to be that the WA and SA button-presses will be tomorrow. Any predictions?

    I’m sticking with the leaders – GRN to take the 12th seat in WA and ALP in SA. I don’t think either will be close – expecting at least 0.1 quota between 12th and 13th in each case.

  20. 1.1 right minor preferences is sizeable. I would not be surprised to see the Greens miss out in WA.

    SA looks a bit more clear for the 4th Labor, but Family First should leapfrog One Nation after Liberal preferences, One Nation along with minor right preferences could make it close, but I still expect Labor to get the last seat (though I won’t be shocked if Bob Day gets there)

Comments are closed.