Slicing up the apple: episode two

It really is a shame that the South Australian election has had to take priority in the past few weeks, because a Tasmanian election is a gift that keeps on giving. For the past nine days, fortunes have continued to wax and wane with each elimination and distribution of surpluses. The process has been of particular interest for the Greens, whose election night fear of two lost seats has faded with the progress of counting in Lyons, and who may yet hold their seat in Bass. If so, they will have returned all their sitting members and a great deal of the initial post-election commentary will need to be revised. The count has been less kind to the Liberals who have been the losers out of the Greens’ strengthening position in Lyons, and are watching their second candidate’s lead in Franklin get slowly whittled away by Labor. If that continues at the expected rate, they will have registered no improvement on the 2002 disaster.

Of the five electorates, only Denison (3-1-1) and Braddon (3-2) have never been in doubt. In Lyons, Labor comfortably returned three members while Opposition Leader Rene Hidding was the only Liberal to safely make it home. That left the second Liberal candidate, Geoff Page, playing catch-up with Greens incumbent Tim Morris as counting progressed. However, Morris’s lead has in fact opened slightly, and it is now clear that the result will be 3-1-1. In Franklin, Liberal candidate Vanessa Goodwin has been fighting it out for the final place with Labor incumbent Paula Wreidt, the other seats having gone two Labor, one Liberal and one Greens. Early on Goodwin was thought to be the favourite, but a complicated conjunction of circumstances has Kevin Bonham of the Tasmanian Times forecasting that her current 259-vote lead will become a 700-vote deficit after distribution of preferences from soon-to-be-eliminated Labor, Liberal and Greens candidates.

In Bass, opinion is divided on whether Greens member Kim Booth can still hope to prevail over the third Labor candidate, Steve Reissig. Booth’s main booster seems to be the aforementioned Kevin Bonham, who has calculated that a "plausible" leak to Booth of 12 per cent upon the elimination of Labor’s stragglers will give him the narrowest of victories. Antony Green has arrived at a higher estimate of the required rate of Labor leakage, and will only say that it "could yet be close". First-rate number cruncher Geoff Lambert does not share Bonham’s view regarding the likely rate of Labor leakage, telling the Poll Bludger that the 10 per cent he has factored into his own modelling is "generous" to the Greens. Lambert says the rate of leakage has been far too low in the most recent round of counting, in which preferences have been distributed from excluded Labor candidate Grant Courtney. However, he also says that "we do not know whose votes Courtney had taken away from him", whereas it seems Antony Green does – he tells Upperhouse.Info that they were Courtney’s primary votes (of which there were 1723), and that the remaining 775 awaiting distribution came to him as preferences. Since not all of these will have been first-preference Labor votes, they will presumably be more likely to leak to the Greens.

Author: William Bowe

William Bowe is a Perth-based election analyst and occasional teacher of political science. His blog, The Poll Bludger, has existed in one form or another since 2004, and is one of the most heavily trafficked websites on Australian politics.

19 comments on “Slicing up the apple: episode two”

  1. The Tasmanian distribution of preferences is conducted under a new act. For the first time, when a candidate is excluded from the count, all votes of equal value are distributed at the same count. Previously, votes were distributed in smaller bundles in the order a candidate received them. That’s why there have only been 30 counts to this stage in most electorates, where under the previous Act there would have been several hundred.

    The exclusion of Courtney’s vote would first be of all full value votes, then of all votes at lower transfer value in order of descending transfer value. That means almost all of his votes yet to be distributed will be reduced value surplus votes of Michelle O’Byrne.

  2. People should have a close look at the Denison result as well. It is strongly rumoured that Michael Hodgman will retire in the next term causing a by-election. However, because of the way countback by-elections work under Hare-Clark, the second finishing Liberal Fabian Dixon will be at a disadvantage compared to the other three Liberals as he was unexcluded at the point where Hodgman was declared elected. That means that at any re-count, these Liberals get their primary votes back from Hodgman, but Dixon will not because his votes were not excluded at the point where Hodgman was elected.

    The current re-count method gives little weight to the fact that people vote for party as well as candidate. The best way to overcome the problem would be to completely count the votes again but exclude one candidate. This is used in other states where electronic counting is used, but very messy and time consuming in Hare Clark compared to the current re-count method of only examining the candidate’s quota. If anyone can think of a shorter method to overcome the problem, the Tasmanian parliament and Electoral Commission would love to hear of one.

    The current electoral act recognises parties by allowing a vacancy to be filled by nomination in the case where no candidate of a party nominates for the re-count. But that still doesn’t cover the problem Dixon will face where his high primary vote will not assist him in any by-election.

  3. I’d like to add a strong note of agreement with Antony’s second comment. This is a bad theoretical bug in the Hare-Clark recount system that should urgently be addressed. It is not just a party issue but can also disadvantage candidates at local government level, where candidates excluded very late in the count or left unelected without exclusion can be hugely disadvantaged.

    The 12% approx leak required figure I gave was not from Labor to Booth but from Labor to any source including exhaust – any leak from Labor to Booth counts for 2 votes in terms of the margin. If Cripps leaks as much as Courtney did, then I expect Booth will just win since most of the Lib preferences will exhaust and those left over or leaking probably won’t be strongly enough pro-Labor to make a difference. Whether Cripps will leak that much remains to be seen and another note of caution is that Fry is an ultra-conservative whose preferences may be more pro-Labor than most Libs. I am posting updated comments to TT as fresh figures arrive but have no control over when they are placed on the site.

  4. I agree that the bug in Hare-Clark should be fixed (have argued it for years actually). However, I’d regret it if the way they fixed it was going over to electronic entry and counting of votes. The count is so much more transparent when it is done like this than when the votes get entered by data entry staff one cannot possibly keep up with, and then a button is pressed and a result comes out.

    I have not thought it through properly, but might it be possible to do a recount including the votes of the candidate who has resigned, and all votes to unelected and unexcluded candidates at that stage. This would still be more work than the current system, but not as much as starting from scratch and doing a complete recount with that candidate excluded.

    Haven’t thought through the full implications, and whether this would create new and larger bugs.

  5. Wriedt just elected – has 308 votes more than Allie and Goodwin combined, so that’s the end of the ball game without any need to throw Allie (not sure if they do so to get her to a quota for potential recount reasons or not).

  6. Well, well, well. Second election in a row that Labor’s lack of a third high profile candidate in Bass has cost them. The exclusion of the last two Labor candidates saw considerable leakage of preferences. From Labor’s total being 538 ahead of the Green, the Greens now lead by 400-odd, with only about 1,700 Liberal surplus votes to come. Looks like the Greens will live on with Kim Booth in parliament and official party status.

  7. Cripps excluded – Booth leads Reissig by 425. Looks fairly good for Booth now but I want to do some more projections to see if Reissig has any real chance to catch him off Fry. I recall that in past cases where the last Lib went out (which will happen once Fry has elected Gutwein and Napier) the bulk has gone to exhaust and the rest has not split strongly.

  8. The problem for Booth in the end will be the stable government scare campaign. It will serve to a. reduce the exhaust of the 1700-odd excess and b. increase the proportion of prefs to ALP. The ALP only need to claw back 425 (net). That’s 25% of the 1700. Surely not too hard?

  9. Yep yep I know. Sorry I should’ve been clearer… 25% NET. Net… the difference between the proportions that go to the ALP and the Greens. Surely not to hard from the Libs…

  10. There are two comparable recent cases in which a Liberal was excluded from the count and all remaining Liberals elected as a result, leaving the remaining votes to split between Labor, Green and exhaust.

    In Denison 1998, 2861 votes split Labor 683, Green 541 and exhaust 1637, for a Labor gain of .049 votes per vote thrown (about 1 vote in 20), or a 56:44 split Labor/Green of those not exhausting.

    In Franklin 1998, 2867 votes split Labor 888, Green 640 and exhaust 1339, for a Labor gain of .087 votes per vote thrown (about 1 vote in 11), or a 58:42 split Labor/Green of those not exhausting. However there were two Labor candidates still in the count so that figure is inflated compared with Bass.

    This election, 1716 votes will be split between Labor, Green and exhaust. Assuming exhaust is the same as in Denison 1998 (57% of votes leaving the Liberals), Reissig needs a 79:21 split Labor/Green of those not exhausting.

    While Fry is extremely conservative, even when he was excluded from Bass in 1998 his leakage favoured Labor over Greens 75:25 when there were four Labor candidates left to one Green and Booth’s profile was way lower than it is now. The split is unlikely to be that drastic this time and while Reissig may still close this is now a very likely Green win.

  11. Fry excluded except for 51 votes. There are now 2314 Liberal votes to throw of which 988 will go to getting Napier over, leaving 1326 for Labor, Green and exhaust. From trailing by 425 with 1716 available, Reissig now trails by 407 with 1326 available and that 79:21 required I mentioned above has blown out to 86:14. Looks pretty much over to me now – Booth by about 300 most likely.

  12. Gutwein surplus thrown – Napier surplus of 1107 to go. Reissig continues to gain but too slowly – now 385 behind. He gained only 40 off the first 609. Based on past elections there is no evidence the votes that have gone all the way down the Liberal line (via Fry, Gutwein and Napier) will behave much differently to the rest – if anything they will exhaust slightly more, because they include those voting 1-5 Liberal and then stopping.

    14-7-4, same as 2002, same distribution in every seat, all 23 recontesting members re-elected, after all that has happened in Tassie politics in the last few years. Amazing!

  13. No, but it is looking very likely. The ALP over-quota is almost certain to go to Morris before the Liberals unless there’s something I don’t understand about Tasmanian politics, like both old parties colluding to destroy the Greens or something.

    Um… er…

  14. Morris has no problems. Even if not a single ALP vote flowed to him he would get up because half the ALP vote at least will exhaust rather than going to the Libs. Add in any leakage from within the Liberals and he’s even further ahead. The only question of any interest at all is whether his margin is larger or smaller than Wreidt’s.

  15. In the latest update Morris leads Page by 1665 votes after rather high leakage from the Liberals’ Jane Howlett. Kerry Degrassi has 4680 votes to throw but 2840 will be burnt up electing Heather Butler leaving only 1840 to split between Page, Morris and the likely recipient of the bulk of these spoils, exhaust. For what it’s worth those Labor votes not exhausting should slightly favour the Greens anyway, so Morris will win by 1500-1900. I was in very little doubt about this one on polling night and none after final primaries. He did extremely well on postals.

Comments are closed.