Friday, May 17, 2019

A few election notes (especially re Tasmania)

Just a few quick notes on a number of things, mainly to bring links to various articles to the top.

Election night coverage

On election night I am very pleased to say I will again be live blogging for The Mercury, starting from 6 pm and probably going til about 11 or maybe later, except for any time when I may or may not need to stop blogging and write a quick article for them.  The link will be posted here when known.  It is possible the coverage will be paywalled; if so there are various subscription offers starting from $12, or it may be there will be a paywall-free link.  Once that has been finished, after a brief break I will relocate to home and then continue on here for a while.


Some links to recent items on my site that may be useful:

How To Make Best Use Of Your 2019 Senate Vote

Tasmanian House of Reps Seats

Tasmanian Senate Guide

Jessica Whelan Situation

I want to re-raise some concerns about the Jessica Whelan situation in Lyons.  Whelan is still on the ballot as a Liberal but is running as an independent after disendorsement (see my Reps guide for a detailed account of the history). There has been a lot of confusion and tardiness about issues such as sign removal and preferencing in the wake of her disendorsement.  Whelan has been spruiking some mostly left-wing policies that may sound appealing while claiming to have moved on from her past history of anti-Islamic and otherwise dodgy social media posting.  (Note that she still strongly denies one post that has been attributed to her, but has produced no evidence that the alleged post was faked.)

My concern is that Whelan claims to have learned and become more informed on these issues, but continues to hang around with Fraser Anning supporters online while still saying that she has strong views on immigration and border protection.  So has she really changed or not?  Not only did Whelan like a post calling for votes for Fraser Anning on the Facebook page of an obscure and obvious Fraser Anning supporter, but she also made a post about foreign aid on her own website that used the same image and icons as one from "Respect Australia Rally", a nasty far-right Facebook site that is supporting Anning and considers One Nation to be sellouts.  (I decline to link directly to that page.)  This sort of usage of skull symbols is a hallmark of Agenda 21 conspiracy theories.

Supporters of far right parties including Fraser Anning's Conservative Nationals and Australian Conservatives have been cheering Whelan on on Facebook and we should be very concerned about who might influence Whelan if she is elected.   I suggest those sharing my concerns about this, or even concerned about having an MP who is very politically confused and not ready for the difficult task of life as a federal Independent, ignore whatever how-to-vote card they are given and put Whelan last or nearly so.

Although Whelan is a bad candidate, it's not clear how widely awareness of this has yet spread, and this situation is highly unusual, especially for Tasmania.  I still expect that Labor will retain the seat, but I don't know that, and you know the old saying ... the trouble with political jokes ...

Tasmania Senate

Details of some bad analysis of the Senate race by the Australia Institute of the Tasmanian Senate race have been released today.  As a general rule I advise readers not to trust any analysis by the Australia Institute on any electoral matters until they have seen me endorse it.  The analysis claims the "Coalition" (in fact the Coalition parties have distinct tickets in Tasmania), the Jacqui Lambie Network, the Greens and One Nation are set to fight out the final two Senate spots, and that One Nation could poll 9.5% in Tasmania.

The Australia Institute's Senate polling has repeatedly had high One Nation votes for Australia generally (recently over 8%) and this is somewhat unlikely given that One Nation's Reps polling is generally around 5-6% and One Nation tends to do slightly worse in the Senate compared to the Reps in seats where it runs in both.  It's possible TAI has the size of their Senate vote right, especially if everyone else has the size of the Reps vote wrong, but:

* If anything, other pollsters (eg Essential which had One Nation on 6.6% for the Reps) are more likely to be overestimating the One Nation vote by not taking account of One Nation not running in every seat.
* There is no history of pollsters underestimating the One Nation vote at recent elections.
* Even if One Nation got 8% nationwide in the Senate (which seems unlikely given competition from the UAP) it would not get 9.5% in Tasmania.  Tasmania is a historically poor state for the party and it got only 2.6% here last time.
* Senate polling tends to inflate the vote of named minor parties, and underestimate those not named.  Pollsters need to make a more serious effort to simulate making a choice between 15-20 parties.
* If One Nation were to sharply increase their vote it would probably be at the expense of Jacqui Lambie, so it's very hard to see how Lambie would be competitive against One Nation if One Nation had 9.5%.

The TAI report (I haven't seen the current edition) warned that the Tasmanian figures should be “treated with extreme caution, as they may have inflated the minor party vote”, so the question is why is TAI releasing this stuff through the media when their own report says it's probably unfit for purpose.

My view remains that the Greens are likely but not certain to win one of the last two seats.  The final seat is hard to predict given the chance of Jacqui Lambie's vote going down, and those in the mix for it would appear to include Labor (either John Short or Lisa Singh), Liberal and Lambie.  Those who could also poll well enough to get into the mix for this seat or who might not get much at all include the Nationals, the United Australia Party and independent Craig Garland.  I don't entirely rule out One Nation but if they do get up I expect it would be off a much lower base than TAI suggests, and it would require a situation in which there is a very even spray of minor party votes and nobody in the mix for the last seat has more than about 4-5%.  This would also require the Labor vote and the Labor preference flow to go down.

Impact of death of Bob Hawke

I expect many people to be asking what impact the death yesterday of Australia's 23rd Prime Minister, Bob Hawke, could have on the result. Some people think that it will increase Labor's on the day vote, perhaps by a large amount.  Others think the idea that it will have any influence (beyond shutting down the rest of the news cycle perhaps) is absurd.

I really can't say anything reliable on this.  Most events expected by the commentariat to move votes actually don't, but there's not been any previous case of such an event just days before election day, and nostalgia is often a force in Australian political discussion.  If Labor now win decisively, it will be necessary to comb through prepoll and on-the-day vote patterns to try to establish whether it was a factor or whether the swings in prepoll and booth voting were similar (which would suggest the polls were overconservative).  However because of the vast changes in the rate of prepoll voting, any such analysis could well be inconclusive.  On this subject, beware of hot takes before all the numbers (including the next two weeks of postcounting) are in.

Messy Reps Seats

I post here a brief list of Reps seats in which it may be unclear who the final two candidates are based on election night totals.   These are seats that may give rise to complex post-counting scenarios and in which two-candidate choices by the AEC on election night might prove incorrect:

Mallee - this could be the messiest; it is two Coalition candidates (though the Liberal is apparently a fly-in) versus multiple independents
Lyons - disendorsed Liberal still on the ballot and campaigning while her party tries (somewhat halfheartedly) to shovel her votes to a little-known National instead
Macnamara - likely race between Labor and Greens to get into the top two.  Note that if the Greens are marginally behind on primaries in such races they can sometimes sneak into second using micro-party preferences.  They narrowly failed to do this last time.
Higgins and Kooyong - ditto to above (though limited and unreliable polling suggests the Greens might be clear in second in both cases)
Gilmore - three-cornered and unclear on recent polling which Coalition party might make the final two

Feel free to suggest others.

Senate Count Will Be Slow

Please be aware that the Senate count will be slow and it will probably only be possible to get some rough ideas about the last few seats late on election night.  In particular in Tasmania and NSW with significant below-the-line uprisings we may have to wait several days to judge the impacts of below the line votes on the contest.

Contrary to popular myth, in the initial primary count all votes (above the line and below the line) are counted, but they are initially counted just as unapportioned and later sorted into above and below the lines for particular candidates.  So there is no big rush upwards in the Green vote (for instance) as below the lines are added - they are already in the mix from the start, and any tendency for the Green vote to rise slightly late in the count (if it does) will be caused by other reasons - later postals, absents etc.

As a result the vote tally for below-the-line candidates like Lisa Singh and Jim Molan will take some time to be clear, and will be very low when we first start seeing figures for them, and will then grow over a period of many days.  Fortunately in the case of Singh we should be able to project her final below the line vote quickly by using her 2016 below the line vote in booths, but for Molan this will be more difficult because of the greater scale of his BTL campaign compared to in 2016.

I may add other notes later today.


  1. Potentially messy reps seats or non-classic:

    I expect Durack to go Lib-Nat but could be wrong on this.
    Is Indi clear? Could be Lib-Nat or Lib-Ind I think?
    Greens could rise in Richmond. Mackellar and North Sydney could also become Lib-Greens but no chance of the incumbent losing.

    Could be messy in Qld. Herbert if either major party drops. Fairfax also has to be some chance. Hinkler is another.

    Wright could have PHON second but don't anticipate that will cause any real problems.

    And in WA can that independent in Curtin get above the Greens and/or the ALP?

    1. Thanks for those. Indi I expect the independent to easily make the final two.

  2. Hi Kevin,

    Getup just sent me an email mentioning an article in The Courier Mail from Brisbane yesterday (16 May 2019) stating that Dutton will win Dickson by 941 votes:

    "A shock new poll shows Peter Dutton on track to win by just 941 votes.1

    [1] Shock result possible in Dickson after massive resources thrown at seat, The Courier Mail, 16 May 2019 "

    I looked on the Courier mail site, but I couldn't find the article.

    Given your views on seat polling, what do you make of this?

    In my opinion, the uncertainty around this "result" would probably be plus or minus 1000 votes!!!


    1. A Galaxy seat poll had Dutton ahead 51-49 so I expect someone has translated that into votes, and doing so to the nearest vote is quite silly. I'd say plus or minus thousands given the poor record of seat polling - if Galaxy are no more accurate than in 2016 then that 51% for Dutton is +/- 5.5%.

  3. Maybe it's too early but we cannot avoid the question of Bob's passig. What will this mean for the election?

    Having campaigned on the ground in Dickson for the past 6 months for Labor I was not too surprised by the 2PP seat poll. I've made more than 700 phone calls in that time and have spoken to maybe 200 people. It's tight, too tight for anyone to call this seat on the night.

    Very unscientifically I have a hunch that Bob's passing will see about 0.5% of undecided voters move into the Labor tent. Perhaps they might not have otherwise. That's about 500 votes in a seat like Dickson. Peter Dutton won by just 1,600 votes last time. The Ali France campaign in Dickson has made 90,000 phone calls in the past 12 months and those phone calls matter right now.

    Peter Dutton is liked in Dickson. Queenslanders are a parochial group of people and Dickson more so than most city edge suburban/rural seats. Peter has spent more than $1million on this campaign and has the full weight of Murdoch press behind him with full front page coverage a number of times.

    I expect a full recount in Dickson and I am unable to say that one party or the other will win it.

  4. Thanks Kevin for all your stellar work the last three years.

    I'm trying to plan my movements tomorrow night. If I remember correctly from 2016, there's not much except exit polls before 6:45pm or so, there's a bit of info trickling out by about 7:15 with quite a few booths reporting by 7:45. Does that sound about right to you?

    Are there any districts you'll be keeping a close eye on for early hints about how things are looking?

    1. It's something like that, maybe a bit faster at the later end.

      I'll be focusing heavily on the Tasmanian electorates early in the night - they tend to report pretty fast anyway - and I'll also be keeping a close eye on a lot of the "non-classic" seats (Coalition vs Ind etc) where the early primary figures are often informative about who is or isn't in with a chance.

  5. Re fighting out the final two Senate spots. Under our current rules where we elect 6 at a time and the quota is one 7th, in most States there will be two quite separate contests for 5th and 6th. If the left and the right both get totals over 42.9% they both get 3 elected. Creeping over the boundary to get a bare majority gains them nothing - to steal one from the other side they need to get to 57.2%, which the right fluked in Queensland in 2004 but doesn't usually happen. Unless Labor stupidly gives its preferences under a Group Ticket to the extreme right (in the shape of Family First) rather than the Greens as happened in Victoria in the same year!

    So there'll generally be one contest for the third place on the left between 3rd Labor and a Green and another on the right between 3rd coalition and someone from the more extreme right. But, without entirely endorsing the AI's claim as to Tassie, it does get a bit more interesting when you get someone who could arguably pinch votes from left and right like Ms Jacqui. But I'd still tip one contest between 3rd Labor and Green, and another between all the rest. The Saturday count will be fascinating and the final count when the date entry is complete will be even more so.

  6. Will the pre poll votes change the outcome - 7.46m pre polls and 16.7m voters ! Could the AEC and Antony Green be wrong !

    1. The projections being done assume the swing in prepolls will be the same as last time. Because there are a lot more of them this may change but there wasn't much change in polled voting intention in the campaign.

  7. Kevin, there are a rising number of people like me who have become entirely sick of the political bullshite and now treat the endlessly repetitive polling as an opportunity to express our disgust. Our responses are all over the place deliberately, with the intent of making the act of polling worthless. The hope is that instead of attempting to work out where to throw the pork ti bait the greedy and self interested few that determine the outcomes of elections lately, our aspiring leadership might actually revert to developing proper policy platforms and start interacting with the electorate as if we were intelligent individuals capable of discerning the difference between wild honey and their core and non core promises.

    Expect the problem of incompetent polling to grow as long as our political class continues to serve anyone but their constituents.