Sunday, May 22, 2022

2022 House of Reps Summary Page And Vanilla Postcounts

 ALP WINS WITH MAJORITY (FLOOR MAJORITY UNCONFIRMED)

SEATS APPARENTLY WON LABOR 77 COALITION 58 IND 10 GRN 4 KAP 1 CA 1 

Seats apparently changing (not all completely confirmed):

COALTION TO LABOR: Reid, Robertson, Chisholm, Higgins, Boothby, Pearce, Swan, Hasluck, Tangney, Bennelong
COALITION TO GREENS: Ryan, Brisbane
COALITION TO INDEPENDENTS: Wentworth, Mackellar, North Sydney, Curtin, Goldstein, Kooyong
LABOR TO GREENS: Griffith
LABOR TO INDEPENDENT: Fowler



In significant doubt:

(Richmond, Lingiari, Macnamara, Gilmore assumed Labor retain)
(Lyons called Labor retain)
(Cowper and Bradfield Coalition retains)
(Menzies, Sturt, Moore assumed Liberal retains)
(Deakin assumed Liberal retain - not yet absolutely confirmed)
(Bass called for Liberal)
(Bennelong assumed Labor gain)

------------------------------------------------------------------------------------------------------

This is the page where I post my summary assessments of the House of Reps count and I will be unrolling details below on seats of Coalition vs Labor interest progressively through the night and updating the post as I go.  Seats will be listed in alphabetical order but will be unrolled on an ad hoc basis.  When I have unrolled them all a Senate thread will be posted; for now see my Late Night Live assessments.  Each post will be updated until I consider the seat of no further interest, at which point it will be moved to the bottom of the page and noted as an assumed win.

Links will also be posted here to non-classic contests or messy contests as I add them:


Note re Ryan: This seat is being realigned to LNP vs Green causing a misleading 2CP count to appear showing the LNP ahead.  That count is just of postals received so far which favoured the LNP, but based on the Greens' preference share in even that sample, the Greens project to 53.2% off other booths.  That cannot be pulled back on postals (especially since the Greens do well on absents); the Greens have very clearly won the seat.  (Update Wednesday 5:30: earlier today I projected the Greens' current lead to be 52.5-47.5 not the 52-48 of the ABC (it is underestimating the flow to Greens because its preference sample is from strong LNP vote pools) - the lead will come down on postals but not by enough to stop the Greens winning.) Note Thursday: Julian Simmonds has conceded.

Note re Curtin: Celia Hammond has cut heavily into Kate Chaney's lead on postals but unless something very odd happens with absents and dec prepolls, Chaney will still win as absents at least will favour her.  There are probably only enough postals left to cut the lead to say 600 anyway (and that is if the flow doesn't weaken in late postals, which it probably will.)  Note Thursday: Celia Hammond has conceded.

Note re Wannon: This seat is being realigned to Dan Tehan vs Alex Dyson (IND) but Dyson needs 85% of preferences to win and this is not possible with too much of the count consisting of right wing microparties.  It might be close but that's all.  After five booths the ABC projects 56% to Tehan; I project 57.3.  These are rubbery projections - it might be 54 or 58, anyway Tehan will win.


Gilmore (NSW, ALP 2.6%)

Gilmore is the seat at most risk of falling from Labor to the Coalition (Labor's only two losses so far have been to the Greens and an independent).  So poor is the Coalition's overall performance that it takes a poor campaign in 2019 followed by a great candidate in 2022 to make it competitive in this seat, which echoes Lyons where there was a disaster in 2019 followed by candidate problems for Labor in 2022.  Even so, that's only been enough to put it on the wire, with former state Minister Andrew Constance (whose state seat of Bega fell in a by-election after he resigned) leading first-term Labor MP Fiona Phillips by 48 votes.  As a recap of 2019, there was preselection acrimony firstly involving Grant Schultz's challenge to Ann Sudmalis, then following the parachuting in of Warren Mundine, and Schultz ran as an independent. As a summary of 2022, Constance became famous following his emotional response after the 2019-2020 summer bushfires.  

Gilmore is at 81.2% of enrolment counted.  In 2019 it reached 92.9%.  In that year absents were 2.3% of enrolment and 5 points more favourable to Labor than ordinary votes, postals were 5.1% and 7.9 points more favourable to Liberals, and declaration prepolls were 3.6% and very similar to ordinaries.  This year there is a larger postal count, with 13.3% applying for a postal and 12020 postals (9.8% of enrolment) already received, though 841 have been rejected.  If postals break anything like 2019 Constance obviously wins (barring miscounts in any booths, which needs to be checked for).  

The postals so far have broken to Constance, but the break is weak, just 4 points more favourable than the ordinaries.  That is enough, but if the late postals/early postals dynamic seen in previous elections persists, then it might weaken further.  I am not sure that dynamic will persist given the shift in apparent voting intention seen in late polling.  So there is some sort of argument that Constance should win by a few tenths, but it is debatable, very close, and could be overturned if there are count corrections.  I think with this one I will put the projection away for the time being and wait to see more votes.  Updates to follow; we could be following this battle for a while.  

Monday: There has been a massive addition of postals, such that there might be only 2500 or so still to come.  Constance now leads by 360 (50.15%).  The break on postals has stayed about the same, around 4.5%.  

Except that Constance has a problem.  In the booth of Gerringong, it strongly appears that the gains on preferences have been entered the wrong way around.


There is no way Constance is getting nearly 80% of preferences in a substantial booth with a decent sized Greens vote.  It seems the numbers (247 and 69) have been credited to Gilmore and Phillips the wrong way round and in fact, Constance's lead is not 360 votes; it's 4.  

[Edit: And Constance's lead has dropped another 54, so he is effectively now behind.]

Tuesday: Gerringong fixed, Constance now 105 ahead but there are now probably more absents left than postals so that lead is very fragile.

Wednesday: All the currently held postals have been thrown bar a few hundred (this may have happened yesterday) and Constance is only leading by 114.  From projections last time I'd expect him to drop about 250 on absents, 50 on dec prepolls, 50 on provisionals and then there are the mysteries of COVID votes.  As against that, maybe he gets 50 back on postals but he is currently projecting to nearly 200 behind, outside the automatic recount margin.  He therefore probably needs absents or dec prepolls to be unusually strong for him to win.  

Thursday 3 pm: Small gain on rechecking for Constance (but it's substantial in the context of a close one); he now leads by 157.

Thursday 4:30: Constance doing exceptionally well on rechecking today and now up by 241, putting the seat back to recount territory on projection.

Thursday 5:40: More gains and lead now out to 307 - not sure if there is anything underlying this, he will be back to favourite if it continues.

Friday 1:00: While there are only nine booths still to be rechecked this includes four large PPVCs so there is still ample scope for a shift in either direction in that process.

Friday 7:20: Two PPVCs to go and Constance's lead down to 250 now.   Hopefully we see some absents soon to clarify where this is going.

Saturday 5:20: Rechecking over and Constance's lead is down to a rather shaky 214.  

Monday 12:10 HUGE BREAK TO LABOR ON ABSENTS.  A batch of absents breaks 334-145 to Labor with provisionals 95-63 and dec prepolls 378-288 putting Phillips 142 ahead.  Unless remaining absents/dec prepolls are from somewhere Liberal-friendly it will be hard for Constance to come back from here.  Still to come up to 1850 absents 3970 dec prepolls, a few provisionals, probably several hundred postals.  

Tuesday: A further batch of absents has broken 306-294 to Constance reducing Phillips' lead to 130.

1:20 Dec prepolls have broken 425-392 to Phillips, and COVID votes have broken 181-122.  Phillips firms further and now leads by 222 with perhaps 1200 absents, 2000 dec prepolls and a few hundred postals to go.  Reports that Labor has claimed victory.

5:55 Phillips 242 ahead now. 

=================

Bennelong (NSW, Lib 6.9)

Bennelong, the seat John Howard held for so long then famously lost, was held for one term by Labor's Maxine McKew and then held for four terms by former tennis star John Alexander.  On Alexander's departure Labor selected a star candidate Jerome Laxale and at the moment it looks very much like Laxale is up.  The seat is one of many where the Morrison government's sabre-rattling on China apparently hasn't gone down well with the Chinese Australian community.  With 75% counted Laxale is ahead by 1749 votes.  Absents (with the usual bonus to Labor, in this case of 4.7 points) weren't a big thing in this seat in 2019 and the postal count is also relatively modest compared to some.  In 2022 postals were 10 points more favourable to Alexander than ordinaries but so far in this count the gap is only 6 points. Even if it increased to 10 points for the remainder I would still have Laxale winning by about 600 votes meaning that it's hard to see him not winning without a counting issue.  And on that, I've checked and not seen any obvious outliers.  I have put Bennelong at the top of the list temporarily for aesthetic reasons but it will be moved to the bottom tomorrow unless a meteor strikes Laxale's vote count.  

Deakin (Vic, Lib 4.7)

Michael Sukkar has a controversial term facing accusations of involvement in branch-stacking and investigations of a donor with foreign influence risks.  He is under the pump here against Labor's Matt Gregg and is currently 1077 votes behind.  In 2019 absents were 4.4% of enrolment in Deakin and Sukkar underperformed on them compared to ordinary votes by 4.8 points.  Declaration prepolls were 5% and had little influence and postals were 12.1% and he overperformed by 6.8 points.  Sukkar is off to a good start on postals with a split 8.4 points better than ordinaries in the first lot of postals, but I would not rely on that to hold up.  If the postals follow the 2019 pattern there might be enough left to put Sukkar a few hundred ahead but the absents and declaration prepolls would more than cancel that out and put Gregg a few hundred up himself.  However, that's very close and at this early stage of counting it's hard to be sure there are no counting errors (though I just ran a primary vs 2PP crosscheck on them and didn't find any).  So Labor better placed, but not by much.

As a general comment about postals, it appears that there may be a higher share of those issued being returned than normal.  I'm seeing this across numerous seats and it could help the Coalition save some.

Update Monday: More postals have been added closing the gap to 846.  These were actually very slightly better for Sukkar than the first lot so there is still realistic hope that he can outperform the 2019 projection and catch up; it's not easy and could be very close.

Monday 4 pm: Massive flow to Sukkar in a further lot of postals that has put his advantage on postals up to 10.8 points compared to ordinaries.  Sukkar is now more or less even with a lot more postals to come than absents and now looks more likely to win.

Tuesday: The killer postals don't stop coming and with Sukkar 730 ahead with several thousand of them still to go, I've seen enough of this (so I suspect has Gregg).  Sukkar retains.

Friday: I've been ignoring this count but it seems there's been a remarkable turnaround - the killer postals did stop coming (almost as soon as I said it) - Sukkar is now 655 ahead and there are only 891 awaiting counting (and presumably not many more to come beyond that).  Current postal count is 9319-7158 to Sukkar.  So far absents and dec prepolls combined have done nothing; there are slightly more absents than dec prepolls remaining and it would not surprise me if the flow on absents to Labor strengthened to 55-45 or so.  Even so there just don't seem to be enough votes left unless there is a counting error.  There are still some rather large prepolls awaiting checking so to be cautious here I am putting this one back into the significant doubt range, though Sukkar should still just retain.  

Monday: Sukkar's lead out to 887 after a batch of postals decided they liked their old stuff better than their new stuff and broke 849-617 to him.  We can probably go back to sleep on this one now.

Monday 2:30 The huge Ringwood prepoll is the only booth yet to report on rechecking.

Monday 3:00 A big pickup for Labor on absents cuts Sukkar's lead back to 613 [edit: 604], however there are only 1171 of those left which might take it back to 500 or so.  Ringwood PPVC still not rechecked.

Tuesday 2:00 Sukkar lead out to 663.  Clearly not enough votes left without counting errors but still awaiting evidence that Ringwood PPVC has been rechecked.

3:30 Lead 608 after COVID votes.

Lingiari (NT, ALP 5.5)

Lingiari was considered vulnerable because of the retirement of veteran Warren Snowdon and because one never really knows what happens in the nation's most unpollable seat.  Marion Scrymgour (Labor) is 1043 ahead.  The issue for the CLP is that there are really not enough votes left.  There may be just over 1000 postals left, in 2019 there were about 1000 absents and about 2000 declaration prepolls.  As a result of logistic issues, turnout appears to be falling from 73% to something like 65%.  There is also a correctional centre booth not showing, but even if it has votes they will be massively pro-Labor.  I have put Lingiari here but will move it to the bottom of the article later today as it appears Labor has clearly won it.

Lyons (Tas, ALP 5.2%)

Lyons is Tasmania's central rural/regional seat riddled with tiny towns, which votes somewhat similarly to Bass and Braddon with a few points extra for Labor.  In both 2019 and 2022 Lyons has seen controversies regarding candidate posts on social media.  In 2019 these led to the disendorsement of Liberal Jessica Whelan and the Liberal Party had to endorse the Nationals instead (all this after Whelan's name was on the ballot paper).  In 2022 the revealing of some fairly old (pre-political-career) social media posts by Brian Mitchell resulted in him deleting his social media accounts and being targeted in Liberal attack ads (somewhat hypocritically given the Liberal Party's issues with women in this term.)

In ordinary votes, Mitchell has a 50.43-49.57 lead over his Liberal opponent Susie Bower.  In 2019 the shift in postcounting reduced Mitchell's lead by 0.23 points.  Lyons has a greatly increased postal count in 2022 with 10347 postal envelopes returned already compared to 6500 admitted to the count in 2019.  640 postals have been rejected already, but there are probably still at least 1000 to come.

In 2019 postals favoured Whelan over Mitchell by 2.3 points compared to ordinaries, and if this is projected to the 2022 Lyons count then the seat becomes extremely close.  However, there is a pattern in the Lyons votes from 2019 that a higher share of the Coalition vote is for Whelan in postals and prepolls compared to the booth votes, reflecting growing awareness among booth voters of the scandal affecting Whelan and her disendorsement.  In 2022 it may be the reverse: as the campaign went on, more voters were exposed to attacks on Mitchell.

This is certainly born out by the first batch of postals, which have broken to Mitchell 2056-1833, making his 2PP share of postals 2.4 points better than ordinaries.  If this continues, Mitchell wins the seat and his current lead of 704 is already hard to overturn without errors in a state where postcounting breaks weakly.  However, there is also some potential for the late postals to be bad for Mitchell for the same reasons as above.  Nonetheless it currently looks likely Mitchell will retain.

UPDATE 1:45am 

In comments Anthony Llewellyn wondered whether the postal count was irregular in Lyons because Mitchell was getting most of the postals.  Having run health checks on some of the other counts (remember Wentworth 2018?) I thought I'd do the same for Lyons. The postal count was fine but one of the others is not!  


Two booths are big outliers.  MC is Mole Creek, a smallish booth with a high One Nation vote.  But LPPV is Launceston Lyons PPVC, a large prepoll with 3469 votes and a bog-standard distribution of minor party primary votes.  The current figures have Bower gaining 626 votes and Mitchell gaining 358.  I believe that what has happened here is the numbers have been added the wrong way round and in fact Mitchell gains 626 and Bower 358, or some other error of similar magnitude.  (An alternative is that the 2PP is right and the primary count is wrong in some way but that seems less likely.)

If it is what it appears to be (a common problem) then when fixed Mitchell's lead will increase by over 500 votes placing the seat beyond what doubt remains.

Update Monday 11:35 Another ball bounces in Mitchell's favour with correction of an error in the Midway Point booth pushing his lead out by another 100 votes or so.

Update Tuesday: Another c. 2000 postals have been counted breaking in Bower's favour and taking Mitchell's official lead down to 719.  

Wednesday: Mitchell's lead is now down to 535 with 3580 postals awaiting processing.  Postals after the first batch have been breaking to Bower 54-46, which if it continues could knock Mitchell's lead down to say 200, leaving him at the mercy of dec prepolls.  But the Launceston Lyons PPVC booth which appears to contain 536 votes of buried treasure (see above) is still unchecked, and if I'm right about that, that would still put Mitchell's win beyond doubt.  

Thursday 2 pm: The Launceston PPVC issue has been fixed putting Mitchell's lead over 1100. Called.

8 pm: Remaining postals have knocked Mitchell's lead down to 784 but there are not many postals left.

Friday 5:49: Rechecking has knocked Mitchell's lead down to 603.  The remaining big booth to be rechecked however is Sorell PPVC where the swing is substantially worse than the Sorell on the day booth so I would not expect much damage in that one.  Incidentally the swing on prepolls is almost exactly the same as the swing on the day in Lyons, suggesting that absents and dec prepolls will probably behave much as they have before.

6:40: The first absents put Mitchell's lead out to 678.

Monday 12:10: All the ordinary booths have been rechecked without incident with Mitchell's lead out to 688.

Menzies (Vic, Lib 7.0)

Menzies is a vacancy for political veteran Kevin Andrews who lost preselection to lawyer and Army veteran Keith Wolahan.  Wolahan looks like he should succeed Andrews, but only just.  Wolahan had 50.03% in the ordinaries and in 2019 the postcount gain was .49.  Postals were 7.2 points better than ordinaries last time and are 7.5 points better so far this time, and while that may very well contract somewhat Wolahan should still win.  Since his margin is just 925 at this stage I am waiting for more votes.  

Update Monday: Wolahan up by 1744 now and has retained the seat for the Liberals.

Moore (WA, Lib 11.6)

Moore is another WA seat that has seen a fearsome swing but Ian Goodenough seems to have done just enough to survive it.  On ordinary votes he is ahead 50.33-49.67 meaning Labor needs a postcount shift of .33 to beat him.  In 2019 the shift was .26 the other way.  Postals so far are 7.8 points above ordinaries to Moore, which is slightly stronger than last time, and has already put Goodenough 1138 ahead.  I see no reason to doubt Moore will be retained.

Sturt (SA, Lib 6.9)

Alongside the demise of Boothby, Sturt was now and then mentioned as a Liberal seat at risk, especially after One Nation recommended preferences to Labor. It looks like James Stevens has survived and the numbers are almost identical to Menzies.  Stevens got 50.05% in the ordinaries and in 2019 the postcount gain was .43.  Postals were 7.7 points better for the Liberals last time and are 7.8 points better so far this time.  Stevens' margin is currently 998. 

Tuesday: Stevens margin now 1261. Pretty sure that's enough now.

Friday 10 June: Remarkably at this late stage with the seat having already been declared, an apparent booth 2PP error has been spotted by one of my Twitter followers in the booth of Beaumont!


(The other outlier is a hospital booth with 45 votes).  This looks like the same thing as the other errors graphed above, ie 2PP gains for Liberal and Labor entered the wrong way around.  If so, it takes 468 votes off James Stevens' current 1481 vote lead.  He still wins, but drops inside 50.5% with a few hundred votes to win, meaning Sturt is very marginal.

11 comments:

  1. Looks like the standard old blaming of minor parties for a major party's defeat has already started. Not sure if you've seen it, but this time it looks like right-leaning commentators complaining that the Coalition wasn't "right-wing" enough and lost voters to One Nation/Palmer's UAP/Liberal Democrats.

    ReplyDelete
  2. Hi Kevin, I'm struggling to understand why Wannon is not listed as in doubt by anyone. Can you help me with that? The 2PP count is Lib vs ALP, but the Independent is clearly going to finish second, and looks, to me at least, to have a chance of winning.

    ReplyDelete
    Replies
    1. The preference flow required for either Labor or the Independent to win is around 85%; this will probably increase with more counting but in any case is impossible given that there are right-wing micro parties whose preferences will help the Liberals.

      Delete
  3. I've been following Vic senate with interest. Not sure if there have been some postals coming in but seems like Labor and Liberal have grown their shares marginally. We have:
    UAP: 0.2968
    ONP: 0.2030
    Legalise Cannabis: 0.2338
    Labor: 0.1930
    LDP: 0.1676
    LNP: 0.1418

    Seems like LNP must be an okay chance of getting above LDP and snowballing. There is also around 0.45 quota of smaller left/centre-left/centre parties that will go out before Labor (Greens, socialists, Fusion, Reason, Dems, DHJP), as well as a chunk in Shooters etc. so Labor must be a very good shot of getting ahead of Legalise Cannabis for a decent snowball of their own. They probably need LNP to get excluded before ONP/UAP maybe?

    ReplyDelete
  4. Is Curtin still solid for Chaney? Postals are overwhelmingly stronger for Liberals and Chaney only leads 50.9/49.1 at the moment.

    ReplyDelete
    Replies
    1. Yes unless something very weird happens with absents and dec prepolls. The flow on postals is weakening slightly as it nears the tail end and I don't think there will be enough to get it below 500 - absents will help Chaney, dec prepolls shouldn't hurt her much since she did OK on the regular ones.

      Delete
  5. This comment has been removed by the author.

    ReplyDelete
  6. Kevin, I note the ABC still has Deakin in doubt and Sukkar's lead has reduced since you last looked at it. Any prospect of a boilover?

    ReplyDelete
    Replies
    1. I wouldn't think so on current numbers as there don't seem to be enough votes left; even the remaining absents breaking 55-45 would not be enough. But the postal flow to Sukkar has slowed down remarkably. Maybe if there are counting errors it could still happen.

      Delete
  7. Deakin is a Coalition seat at risk to Labor (best case for Liberals is a hold not a gain), not a Labor seat at risk to the Coalition.

    ReplyDelete
    Replies
    1. Ta, fixed. Under massive time stress this election and the blooper count is rising.

      Delete

The comment system is unreliable. If you cannot submit comments you can email me a comment (via email link in profile) - email must be entitled: Comment for publication, followed by the name of the article you wish to comment on. Comments are accepted in full or not at all. Comments will be published under the name the email is sent from unless an alias is clearly requested and stated. If you submit a comment which is not accepted within a few days you can also email me and I will check if it has been received.