At the end of election night, Labor had definitively won 45 seats, but was expected to pick up enough extra seats to win a majority (requiring 47 seats). As of Sunday lunchtime they were leading in 49 seats. This had dropped to 47 when I was analysing the results yesterday, but it looked like there were 3-4 others where Labor was in a good position to catch up.
As yesterday afternoon and evening stretched on, there were a number of updates which were bad news for Labor and seemed to weaken Labor’s position in some of those close races.
By the end of the night, Antony Green announced that Labor would likely fall short of a majority.
Looking like Labor will finish with 45 or 46 seats.
— Antony Green – elections (@AntonyGreenElec) March 27, 2023
So for today’s post I wanted to run through all of these close races to understand what has happened.
The short answer appears to be that Labor has generally gained smaller swings in the pre-poll vote than the election day vote. This is a challenging area for election prediction – the principle of booth matching can deal with a situation where one party does better on one type of vote than another, but it doesn’t deal with a situation where that gap changes, particularly when that situation is repeated in numerous seats. This can lead to incorrect projections piling up, rather than cancelling each other out.
It seems like the ABC was correct to not call any of these seats despite the Labor leads on the night, but it is true that the narrative did settle into an expectation of a Labor majority. I did try to be cautious but perhaps not cautious enough.
I’ll run through these seats in a rough order of which ones seem competitive for Labor.
At the moment there’s a few of these seats where I’d like to see more before coming to a conclusion, but it seems most likely Labor will end up around 45 seats, with Ryde as their best prospect to win a 46th seat.
Kiama
If you calculate 2CP swings in Kiama compared to the Liberal vs Labor 2PP in 2019 (conveniently also including Ward), you get a swing of almost 14% on the ordinary votes. Although that doesn’t yet include Bomaderry Public, a relatively large booth with a 42% Ward primary vote (compared to 37% across all ordinary booths) so that will probably shrink slightly.
Yet Labor only gained a 6% swing at Albion Park pre-poll, and an 11% swing at the Kiama pre-poll. They did much better at the election manager’s office, but that is much smaller. We’re also still waiting for the Nowra pre-poll, but that was Ward’s best pre-poll booth in 2019, with a 2PP of 69.2%.
We also don’t have any postal votes in Kiama – Ward polled over 70% amongst postals in 2019.
Ward still trails Labor by 752 votes, but it looks likely he will win the remaining pre-poll votes and the postal votes, which likely puts him in front.
Having said all that, I can see why people think Ward will win Kiama but I don’t think it’s decided yet.
And just as I was about to publish this post Antony posted an update from Kiama that is not yet in the voting data:
Waiting for the official count to arrive but scrutineer reports are that Ward has won the Nowra early vote centre by 1,200 putting him 600 votes in the lead in Kiama. #nswpol #NSWvotes
— Antony Green – elections (@AntonyGreenElec) March 27, 2023
This should lock in Ward.
Ryde
Labor gained a 10.9% 2PP swing on the ordinary votes in Ryde, but the pre-poll votes so far compared to last time have swung by just 6.6%.
Labor managed 44% of the 2PP in the first batch of postals, which compares to 32% for all postals in 2019. Generally Labor improves their support in later batches of postal votes, so it’s possible that Labor could actually win some later postal vote batches. That is the main reason why this seat is still in play, but it’s going to be a challenge.
Right now Labor leads by just 234 votes, so they’d have to turn around the postal votes pretty quickly to win, as well as doing well on other votes like absent votes.
Terrigal
Labor has been trailing in Terrigal but until recently it looked like they could pick up ground.
Liberal MP Adam Crouch leads by 87 votes, thanks to smaller swings in the pre-poll booths.
Labor gained a 13.9% 2PP swing on the ordinary votes, but the overall pre-poll swing at the moment is just 11.1%, with Woy Woy yet to report.
Labor gained a 12.6% swing at the Terrigal election manager’s office and 11.5% at the Terrigal pre-poll booth, but managed just 3.5% at the Gosford booth. That booth was the one that really changed the picture here.
We are still waiting for the Woy Woy pre-poll booth. This booth was relatively small in 2019 and had a Liberal 2PP of 62.3%. If the booth swings in line with the Terrigal pre-poll booths or the ordinary booths it will end up about tied, but it’s also possible the swing will also be smaller there. Either way, it’s not helping Labor rebuild its lead.
As for the postal votes, Crouch managed 62% of the first batch. Labor would need quite a dramatic turnaround to not be hurt badly by the addition of more postal votes.
You’d want to see the Woy Woy booth and more postal votes before calling this but this is very clearly trending Liberal.
Miranda
There was a 15.2% 2PP swing to Labor on ordinary votes in Miranda but so far we don’t have any pre-poll booths. Liberal MP Petinos managed 59% of the first postal vote batch, and is now leading by 532 votes.
We don’t have any pre-poll votes, but they would have to be very strong for Labor for this seat to be in play. This seat will likely be called for Petinos when some pre-poll votes are reported.
Holsworthy
The 2PP swing on ordinary votes in Holsworthy was just 7%. We also don’t have any pre-poll booths reported in this seat yet. Liberal candidate Tina Ayyad leads by 340 votes, thanks to winning 54% of the first postal vote batch.
We don’t have as much information in this seat. On the one hand, the Liberal is already leading so any pre-poll lead would solidify her position. But on the other hand it doesn’t require such a large swing for Labor to win the pre-poll vote or even the late postal vote, so I wouldn’t rule that out.
We do have primary votes for two of the pre-poll booths and so far they are a tiny bit better for the Liberal than the primary votes on election day. So that suggests they will likely have little to no impact on the margin.
Goulburn
This seat had quite a small swing and the Liberal is leading despite only holding the seat by a 3% margin. Tuckerman currently leads by 306 votes.
Labor did win the ordinary vote with 51.25%, which was a 4.4% swing. But so far, with two out of four pre-poll booths reporting, Tuckerman has actually gained a pre-poll swing, with her better pre-poll booths yet to report. Labor polled 50.6% at the pre-poll booth in the town of Goulburn in 2019 – this time they only managed 48.4%. Meanwhile Tuckerman won Yass with 53.4% and Mittagong with 63.5%, and neither of their equivalents have reported yet in 2023.
One thing worth mentioning, however, is that the election day vote showed a change in pattern. If you compare my pre-election booth map to the post-election booth map in Poll Bludger’s results, you can see that Labor picked up ground in Yass and the Southern Highlands and went backwards in Goulburn. So it’s possible that we may see a similar pattern in the pre-poll vote.
Having said all of that, Tuckerman has also gained a lead in postal votes and that is likely to solidify her position.
Oatley
Oatley looks completely out of reach now, with Liberal MP Mark Coure leading by 910 votes. He won the ordinary vote by 123 votes. There was potential for the pre-poll vote to change things, but he managed 52% of the 2PP for the pre-poll booths, winning the larger booth. He also gained 346 votes on the postal vote.
This seat did look like there was a chance Labor could pick up enough ground on the pre-poll vote to win but not any more.
Will the ALP contact the 3 Independents of Orange, Barwon and Murray ? That would get them over the Line.
At what point do the various ECs bite the bullet and start counting pre polls/postals prior to 6pm?
At some stage, we will go to bed on Saturday night with one mob claiming victory, only for the other mob to get in after the rest of the votes are counted.
They are already over the line with Sydney, Lake Macquarie and Wagga Wagga according to the ABC.
Wakehurst, Wollondilly, the 3 ex-SFF and the 3 Greens were not involved.
Prepolls and whatever postals there are should be counted after 6pm when ordinary votes get counted. I do agree that counting them on Saturday night, instead of the following Monday, would be better and less nail-biting.
Alex Greenwich, Greg Piper and Joe McGirr all confirmed they would guarantee confidence and supply if Minns needed them. I think Minns would reach out to Regan (Wakehurst) or Hannan (a teal from Wollondilly) before the ex-shooters, though I read that Dalton (Murray) is relatively progressive. Even without a majority, Minns has more variety and more crossbenchers to choose from than Gillard in 2010.
Seems as if Libs will have a huge number of marginal seats next time around.
The ex-Shooters MPs were very friendly with Labor in the last parliament. It’s the old adage – “The enemy of my enemy is my friend.” The “enemy” in this case being the National Party. It may well be true to say that they actually have more in common with Labor than the Coalition. They strike me as fiscally left-of-centre (virtually everything they want involves significant spending and/or government intervention) and socially libertarian (they supported abortion but opposed vaccine mandates).
“At some stage, we will go to bed on Saturday night with one mob claiming victory, only for the other mob to get in after the rest of the votes are counted.”
Parties are pretty cautious about this, usually. I don’t think it’s anything to be concerned about. There was no chance on Saturday night and there is still no chance now that the Coalition are going to form government in NSW.
The thing is Ryde and Terrigal were in the gained seats on the ABC website on Sunday morning when it stated they had reached 47.
Agree pre poll should be counted as much as possible on the night. but isn’t the real solution to follow ACT example and have an election week rather than an election day. Lets say Saturday to Friday and count all votes possible on Saturday in lock ups and then have booze up as usual on Saturday night?
Is it really the responsibility of electoral commissions to accommodate the desires of psephology geeks like us?
The LNP Government stuffed up the iVote, that gave direct results at the push of a button, shame it wasn’t used this time.
It is the responsibility of the ec to run a better operation… no 50 to 60 lines waiting to vote. This requires better planning a permanent presence and better training for those who will help on election day.. consider changing the law re postal and prepoll voting
It is the responsibility of the ec to run a better operation… no 50 to 60 lines of people waiting to vote. This requires better planning a permanent presence and better training for those who will help on election day.. consider changing the law re postal and prepoll voting
The ACT has voting on the Saturday just like in every other jurisdiction. What helps is that a large proportion of the pre-poll vote is cast electronically (in person) but that’s a lot easier in an urban centralised polity where most of the pre-polls are cast in a handful of locations.
As for iVote, if they ‘stuffed up’ it was in scaling it up too much. It should have been preserved for people with disabilities or who lived somewhere too remote to vote in another way. It wouldn’t be a solution to the speed of vote counting at that scale.
The iVote stuffed up due to the LG Elections when a huge amount to people were on holidays, and turned up at Polling Booths to Vote and were told they had to use iVote, that caused the system to go i nto overload. Now people overseas were not able to use iVote thus they more than likely were not able to vote in the NSW Election this time.
@Nicholas,
The exact opposite is true. ‘Psephology geeks like us’ have no issue with the long count, it is the damage to democracy amongst the general population that is a concern. We are all on here 3 days after election to follow the count, give our opinions etc, but most people just want to wake up Sunday morning, find out who won, hope it is their ‘team’, then forget about politics to the next Federal election. Most people do not want to find out the results as reported Sunday morning were wrong, and it feeds into distrust, even fears of fraud. The US has a huge problem with whichever side loses claiming fraud, which they can do because the system is so fragmented and poorly run it feeds distrust. We need to avoid that if we are to keep faith in the process.
Agree Labor Voter, many ordinary citizens are fearful and/or concerned when they see results have changed dramatically, even though it is expected with different types of votes leaning in different directions. Having all pre-poll votes counted on election night may assuage some of those fears.
What problem are we trying to solve here? The election night result was reported correctly – the ALP are forming government in NSW. Pre-poll and postal votes have not changed that. Majority vs minority – I suppose if you’re a media commentator or an online pundit (like us!), it’s “important” but the change of government thing is the key here IMO. I don’t think we’ve had a recent (any?) case where late election counting has changed an election night result. I suspect election analysts will change their models in the future to be a bit more circumspect about declaring seat results on election night.
@SP: I don’t agree. It’s just luck that in the last 2 elections the swings have been so strong as to produce a definite result. The system has to be improved so that pre-polls go in on the night. The solution needs to be resource efficient and so I still favour my counting day proposal. On the matter of pre polls I neglected to say that the answer is probably obvious, the conservatives doing a bit better is probably not surprising as the pre poll voters on the average contain a bit more conservative demographic group – older retired men and women.
You don’t need to know the definite result on election night. Is there some kind of rule that says we need to? You can wait a few days or even a week or two. Other jurisdictions cope with this (ones with regular minority governments) and we can too.
The normal pattern is prepoll votes similar to on the day votes.. this appears not to be happening. Normally the postal votes favour the coalition especially in rural areas. All that happened is the close seats less than 1 to 2% did not stay where they were on the night…. the sums for opv held the party with the highest primary vote Labor formed government
Prepolls and postals should be able to be counted on election day before
polls close, with very strict security requirements (Closed room, phones not allowed after scrutineers sign in).
The verification of postal votes could also be done well before election night (I think this already happened).
Opt in computer voting is good for getting a sample of how preferences flow in systems like hare clark, but not sure it’s necessary for single member electorates.
@SP,
No rule, but take off your psephology hat and put on your ordinary bloke hat. You just want to turn up and vote, find out who won and go back to your life. If we keep getting ‘results’ that change a week or 2 after the vote, you stop trusting the process
Labor voter, maybe the fact that ‘results’ are inconsistent is actually an indictment of the single member voting system, because when the vote is so fragmented it is hard to build a majority. Switching to a proportional system or semi-proportional system like NZ MMP may be better so people are geared to accept uncertainty and lack of majorities for each election.
As SP indicated, I’m sure voters in NZ and Europe which have these types of electoral systems are more tolerant and willing to wait when the result is uncertain.
To be clear, the ABC only called 45 seats for Labor. There might have been some chopping and changing through the night but that is the seat count on the screen. Yes there was an expectation the would get to a majority which was reasonable considering the number of close races, and it is unusual how many close races went to the Coalition. But if there was a premature calling of a majority that is the fault of the media narrative, not the psephology.
@Mostly Labor Voter – You won’t get results until the check counts are done anyway. Anything done until that point is a projection and always has been.
That isn’t difficult to explain to the average person. The fact that check counts are done and the official result then comes in a completely open manner should give people that confidence.
We are going to see this happen more and more with elections until we can establish an effective E-Voting process. The US Presidential Election from 2020 is a classic example.
“No rule, but take off your psephology hat and put on your ordinary bloke hat. You just want to turn up and vote, find out who won and go back to your life. If we keep getting ‘results’ that change a week or 2 after the vote, you stop trusting the process”
As I have said, this has literally never happened. You are trying to solve a problem that does not exist. On election night, it was reported the ALP was going to form government. This has not changed – the ALP *has* formed government. As Ben pointed out, even the seat count stuff was reported correctly at the end of the evening. That was a problem with electoral analysis, not electoral process and it was fixed up. There is nothing wrong with the existing vote counting process and forcing people to count more votes after a long day seems not a good thing to me. If the election is very close, you can wait a short period of time for the final result. If the reporting of the result is inaccurate, take it up the relevant media organisation.
Ben, the ABC website did call 47. It was up there at the end of the night and on Sunday. Ryde and Terrigal were in changing seats. Its a consistent ABC issue that they call a few seats too early, I’ve noticed this for a few elections now and one or two don’t end up with Labor. I believe they don’t properly project how mych the prepoll and postals favour the Libs.
Here is part of the late night wrap up on the blog which doesn’t mention Ryde or Terrigal in the seats in doubt:
Sat, 10:55pm
By Leigh Tonkin
Counting has now finished for the night and won’t resume until Sunday.
Antony Green is predicting Labor will end up with 50 seats.
The primary statewide vote is currently at 37.4 per cent for Labor, 25.7 per cent for Liberals, 8.4 per cent for Nationals and 10.1 for Greens.
These are the seats still in doubt:
Goulburn (Labor ahead)
Holsworthy (Labor ahead)
Kiama (Labor ahead)
Miranda (Liberal ahead)
Oatley (Labor ahead)
Pittwater (Liberal ahead)
Willoughby (independent ahead)
Winston Hills (Liberal ahead)
Wollondilly (independent ahead)
In fact it shows three of the above in doubt seats as Labor ahead to make the projected 50 seats mentioned in the post by adding them to the 47 called seats which included Ryde and Terrigal at the time.
Direct link to this is here and you can scroll to the time stamped post to find it: https://www.abc.net.au/news/2023-03-25/nsw-election-live-coverage-blog/102143464
Lnp insider, I read from Anthony green that Labor may end up winning Ryde. This is because absent votes which are yet to be counted will probably favour Labor more than the coalition.
In fact, any of the other 3 seats where libberals lead may have their margins narrow once absent votes are included.
Just a clarification all – I don’t actually disagree with what many of you are saying – and I agree I am overthinking this because no one is really mentioning this in the press. Iin reality Labor don’t even need to get any agreements to govern, they just can. However, I don’t buy the ‘ABC were only giving away 45 on the night’ line, again that means people are paying attention to that and not the rest of the coverage, nor does that mean that other networks (it is amazing the number of people who think only the ABC does election coverage) or press didn’t call more seats for Labor.
All I am saying is it can’t be beyond the wit of man to work out how to count the bulk of pre polls on the Saturday night, we used to get the bulk of votes counted on the night when they were mostly in seat on the day, it can’t be that huge an issue to for example truck them all to a counting centre and have enough people to count them on the night.
On average 3000 to 3500 absent votes each seat. This should be pro Labor and it is possible The later the postal votes the better the Labor vote. Will be does this swing one extra seat to Labor I don’t know
A question for Ben – are absentee votes in NSW more likely to exhaust? In Terrigal, the Libs are 300 or so votes ahead and absent votes are likely to rein this in or give the ALP the lead. But could they lose that advantage with more minor votes exhausting?
I agree with mostly labor voter pre-polls and available postals should be counted by the end of Saturday night. Of course nothing to be done about postals or absents which will be the small number of seats in doubt. Another problem I noticed was the failure to finalise the on the day at some booths. In Kiama this occurred in 3 booths including Bomaderry – a big booth near where ward lives which he ultimately won very well. Levels of competence is falling amongst poll staff because a traditional bulwark of counting – bank teller staff (who could count fast & accurate) no longer exist. Let’s face it, the only people who can count cash fast and reliably now are drug dealers & outlaw gangs who fortunately are busy elsewhere on Saturday nights? The NSW EC needs a big kick up their arses and needs to invest a lot more money in training poll staff.?
I feel that prepoll and postal votes are way less Labor-friendly in battleground seats at this election than at the 2022 federal election. Disclaimer: I haven’t done the maths. Generally, postal voters are more conservative than ordinary voters, but at this election, the addition of postal votes really diminished Labor’s odds of winning a majority.
On the night of the federal election and the following Sunday, nobody said Labor would achieve 80 seats and in fact, it looked like a minority government until the second week after the election.
I remember that Bennelong (where Ryde and Epping are), Lingiari and Lyons were line ball but Labor won with 51% 2PP. Liberals were ahead in Gilmore on election night, I think, but Labor retained it.
I should mention above that the OPV in NSW probably made the difference.
Votante: Lyons & Lingiari sitting members – do more postals work & name recognition helps. Also in lingiari postals tend too be outstations. Gilmore was last addition of pre-polls absents from Albion Park (just outside the seat) but close & probably trades & labourers etc living in Kiama/Kiama Downs but working in the cottage industry in Illawarra – very unusual & not even anticipated by mostly Whitlam booth workers. Bennelong – can’t explain – maybe ghost of frank Walker?
“I feel that prepoll and postal votes are way less Labor-friendly in battleground seats at this election than at the 2022 federal election. Disclaimer: I haven’t done the maths. Generally, postal voters are more conservative than ordinary voters, but at this election, the addition of postal votes really diminished Labor’s odds of winning a majority.”
For prepolls, it can depend on the seat. In East Hills, the prepolls definitely favoured the Liberals – they got a 3% swing towards them vs 4.5% swing towards ALP in election day booth votes. In Heathcote, prepoll and election day booth votes had a similar swing towards ALP – 8.5-9%.
From the numbers I’ve seen, I think you could say generally *in this election* pre-polls favoured the Liberals, but I wouldn’t take that as accepted electoral wisdom going forward.
most of the seats gained at this election are very marginal with the exception of parramatta. i would be surprised if the coalition regains them at next election ballina didnt end up swing to far either. with the right leader campaign and possibly a labor greens governmetn causing disatisfaction i wouldnt rule out the coalition governing in 2027. either in majorty or with a minority plus one or two independants. especially if those former shooter tie themselves too much to labor a majority wont be out of the question
Comments are closed.