Sunday, May 19, 2019

2019 House of Reps Postcount

Coalition has won the election, almost certainly with a small majority

Apparently won Coalition 77 Labor 68 Green 1 CA 1 KAP 1 IND 3. 

Seats Assumed Won By Coalition From Labor: Longman, Herbert, Bass, Braddon, Lindsay
Seats Assumed Won By Labor From Coalition: Gilmore, Corangamite, Dunkley
Seat Assumed Won By IND From Coalition: Warringah
Seat Assumed Won By Coalition From IND: Wentworth 

Seats that were being covered but now assumed won: Boothby and Chisholm (Coalition retains), Lilley, Cowan  (Labor retains), Bass (Liberal gain), Macquarie (ALP retain)

This is the main thread for the 2019 House of Reps postcount.  A few days ago I expected I would be unrolling separate threads to unravel three-cornered contests in Melbourne, a complete mess in Mallee and so on, but at this stage none of that has happened.  I have done a quick thread for Mallee though as there's some interest in it.  There is a weird situation in Hunter, where One Nation are currently two points short of beating the Nationals into second, but I don't see any reason to think they can get into second, let alone win if they do.  Indeed the Nationals' margin in second should increase.

I have been very busy with media commitments regarding the Great Australian Poll Fail and will be unrolling comments on the seats in doubt later this afternoon.  For each seat there will be an intro post and then comments will follow until the seat is no longer of interest.  When the seat is no longer of interest it will be moved to the bottom.  Other seats may be added.  Any seat that goes to a recount will get its own thread for the recount, as may any unusually complex seat that is competitive.

At first the posts will be brief, and more detail will be added when I've got other articles up and running.

Macquarie (NSW) (ALP 2.2%)

Macquarie in outer Sydney often looked like a bit of a trouble spot for Labor and attracted some seat polling.  This looks like being the closest seat with a 625-vote lead to incumbent Susan Templeman (50.37) off ordinary votes but half of that was wiped out in the first 2000 postals.  In 2016 the Coalition gained 0.39% 2PP in the postcount so if that repeats this one will go down to the wire.

Monday 5 pm: Sarah Richards ahead by 23 votes after more postals.

Tuesday 5 pm: An extremely close count with Richards now ahead by 68.  Under 2000 postals likely to be added compared to twice as many absents so Templeman can come back.  It's hard to rely on postcount projections staying the same from election to election so we need to see if either candidate builds a serious lead.

Tuesday 7:30: Richards' lead out to 133, I believe on rechecking.

Wednesday 7:27: Richards has won a batch of absents, but that doesn't necessarily mean much because we don't know where they're from (one cannot project off absents for this reason).  Nonetheless lead now out to 196.

Thursday 11:40: Richards' lead now down to 131.  Another parcel of absents favoured Templeman.  An important thing to watch is whether either candidate can win by more than 100.  A win by 100 or less after the distribution of preferences triggers an automatic recount.  In Herbert in 2016 the recount was fast-tracked.   If the margin exceeds 100 then the petitioner for a recount would need very good reasons.

Friday 8:10: Richards' lead is now down to 71, not sure what on.

Tuesday 6:45: A turnaround and Templeman now leads by 67.  Not all the issued postals ever arrive back, meaning that there could be fewer than 600 postals to go.  There could be over 1000 absents and about 700 declaration prepolls (I'm allowing for likely rejections).  There are 184 provisionals awaiting assessment but most will probably be nixed.  To get back to equality Richards probably needs to break even on absents and prepolls (thus far breaking slightly to Labor) and to maintain the existing flow on postals (late postals often weaken).  So that is somewhat difficult, but maybe she can at least get it within a few dozen, which might then turn around in a recount and/or preference distribution.  On the other hand if the postal split is very weak Templeman might now avoid a recount, but at this stage that would still seem optimistic.

Wednesday 1:30 And another turnaround and Richards is back in front by 7.  This seems to be purely off rechecking as the declaration vote totals haven't changed.

Wednesday 3:00 Labor gets a huge jump on both dec prepolls and absents and leads by 284!  That looks a lot like game over now, it shouldn't even get inside the recount margin.

Monday 1:00: Labor lead out to 298 with less than 1000 to go.  There will still be the distribution of preferences to confirm that Templeman has retained after the check count has finished, but any error large enough to shift it should have been spotted by now.

Seats previously covered but now assumed won:

Boothby (SA) (Lib, 2.8%)

Boothby was the Coalition's most endangered marginal in South Australia, but it helped there that Nicolle Flint had a personal vote bonus as a first-term MP.  Late on the night this seat was projecting to her by a narrow margin.  Currently her lead is 1442 votes.  I assume this is way too much but am waiting to see if there is any reason why the seat is still given as in doubt by the AEC.  In 2016 there was a slight postcount move to Labor in Boothby.

Chisholm (Vic) (Lib, 2.9%)

Chisholm is Julia Banks' former seat, the Coalition's only gain at the 2016 election.  It was a Labor retirement at that election so there is no personal vote effect.  Liberal candidate Gladys Liu (an accomplished former chess administrator, by the way) was widely criticised during the campaign and generally expected to have lost the seat as a result, but it has actually ended up on the wire.  This is a seat in which the Liberal Party controversially deployed mock-AEC coloured signage that Labor has complained bitterly about (restrictions on this sort of thing need tightening.)

Presently Gladys Liu (Lib) trails Jennifer Yang (ALP) by 134 votes, and will gain another 25 or so from a very small prepolling booth not yet added to 2PP.  In 2016 the Liberals increased their vote by 1.11 points after the postcount in Chisholm despite not holding the seat.  Anything like that again will see Liu retain the seat.  Nonetheless the seat was not projecting that strongly before the projections were switched off so let's see what happens here.

6 pm Sunday: Now 169 ahead and a huge postal count to come (a few already in which broke 60-40 to her) so I don't see why Liu will not win.

Tuesday: Liu now over 900 ahead.

Lilley (Qld) (ALP 5.8%)

Labor were at least vaguely nervous about Lilley on Wayne Swan's retirement; I know internal polling was done on it.  The personal vote loss plus the generally poor result in Queensland has fed into a big swing to the government. Anika Wells (ALP) leads with 50.75%.  In 2016 there was a 0.35% swing back to the government in the post-count, so that's close enough to make it worth keeping an eye on.

Monday 7:55: Wells is down to a 901 vote lead now (50.55%) after some unfriendly postals - but not super-unfriendly.

Tuesday 5:00 Wells has benefited from a counting correction and is now up by 1288 (50.79).  There are still maybe 8,000 postals but even at the current break rate those wouldn't be a threat even ignoring other classes of vote.  I'm moving this one to assumed won but will keep an eye on it.

Bass (Tas) (ALP, 5.4%)

Bass is the infamous revolving door seat around Launceston that has evicted the sitting party at 7 of the last 9 elections (easily the Australian record holder) and appears to have done it again, but in spite of Ross Hart's concession on election night it still isn't totally over.  However the Liberals' Bridget Archer is very likely to win.

After all booth primaries were included, Ross Hart (ALP) trails by 437 votes giving Bridget Archer 50.36% 2PP.  In 2016 the postcount in this seat increased the Coalition's margin by 0.78%, but in 2013 when the seat last had a Labor incumbent, the benefit was only 0.04%.  In general Labor doesn't often gain in postcounts, especially not by 0.36 points, but we need to keep an eye on it just in case something happens.  The other possibility early on in any race this close is counting error.

Monday 2:30 - Very, very interestingly the first batch of postals in Bass has broken only weakly to Archer (50.75%) extending her lead by only 22 votes to 459.  Normally the first batch are the worst for Labor!

Monday 4:00 In 2016 postals broke to the Liberals 6.4% stronger than the booth vote, absents the same, provisionals 7 points worse out-of-electorate prepolls 10% better.  But in 2013 when Labor last held Bass. postals were only 2.4 points better, provisionals 6 points worse, absents 1.4 points worse and out-of-electorate prepolls 1.4 points worse.

Even if the 2013 patterns on non-postals hold, Hart would need remaining postals to break to him to win.  Labor did make some effort on postals in the seat, though less than the Coalition.

However a wild card here is the possibility of swing through the campaign in this seat.

Monday 8:00 A further lot of postals has shrunk Archer's lead to 392.  A Liberal candidate has gone backwards on postals.  You read that correctly.  I'm not going to attempt to project this seat for a while because of the weird postal pattern and the history of varying postcount behaviour in this seat. Just sit back and see what the numbers do.  If pretty much anything increases Archer's lead by a couple of hundred or so she should win - Hart can't really afford any type of votes to go against him significantly (which could happen on out-of-electorate prepolls.)

Tuesday 2:40 A new lot of postals has moderated yesterday's weirdness a bit, so all postals so far are now breaking to Archer very slightly and her lead is up to 463.  That makes it a lot harder for Labor.

Wednesday 11:10 Archer lead out to 476, apparently on rechecking.

Wednesday 12:30: A further small increase on postals, Archer now leads by 510.

Wednesday 2:30:  A small decrease presumably on checking, Archer leads by 504.  Assuming postals do nothing from here Hart would need a roughly 57-43 break on absents and out of electorate prepolls even assuming nothing on provisionals.  This difference from the booth count was achieved by Labor in only two electorates in 2016 (Murray and Cowper, both of which Labor was uncompetitive in.)  Moreover, Bass was one of the worst electorates for Labor on this comparison in 2016, and not spectacularly good in 2013 either.  I have moved this seat to "assumed win" status and it will probably be definitively called soon.

Thursday 3:40: With a further slight increase and disappearance of most of the remaining absents I have called Bass for the Liberals.

Cowan (WA) (ALP, 0.7%)

Cowan was one of Labor's big wins in 2016, with Anne Aly taking the seat.  Aly is currently showing with a 0.5% swing to her so a 1.2% lead.  There was a 0.46% move back on post-count in 2016 and just maybe if that is bigger and there is a shift in out of electorate prepolls there might be something here but I doubt it.  Keeping an eye on it for a bit.

Tuesday 5:00  Aly's lead has been cut back on postals to 786 votes (50.54).   Most of the postals are gone now so this points more to her probably having a narrow margin, but it seems doubtful the gap would close completely.

Thursday 4:00: Aly gains on absents to lead by 836 (50.55).  There are still a lot of prepolls to come in this division.

Friday 8:00 Aly's lead down slightly to 786.  Still many thousands of votes to go.

Monday 5:30: Aly's lead is now out to 810.  There are still several thousand votes left but not that many postals and there is no evidence of a meaningful break to the Liberals on any other category.  So I have moved this seat to assumed-won status.  I'll have another look if the gap drops below, say, 500.


  1. Your Poll roundup had herding warning (also thanks to Mark the ballot, that saved me from putting a friendly wager on a large Labor majority. I still used the betting numbers to put money against Abbot.

    In hindsight, given how far out both pollsters and bookies were, I was lucky on Abbot.

  2. You could probably take Wentworth off your 'being monitored' list....thumping results for Sharma on postals seem to have put this one to bed.