At the time of writing on Monday evening, there appear to be 17 close seats worth following over the next few days.
By my reckoning there are three types of seats worth watching:
- Conventional close seats – Bradfield, Bullwinkel, Goldstein, Kooyong, Longman, Menzies, Wills
- Seats now undergoing a fresh 2CP count between clear top two – Bean, Bendigo, Franklin, Fremantle, Melbourne
- Seats where the top two is not clear – Calwell, Flinders, Forrest, Monash, Ryan
For this post I will run through the first category of seats to start with, and follow up with the other categories as time allows.
Bradfield
Nicolette Boele leads by 416 votes. Just over 10,000 postal votes have been counted, but no absent or declaration pre-poll votes have been counted yet.
Based on how the teals polled in each vote category in 2022 (adjusted for the redistribution and Boele’s 2CP swing so far), it suggests she would slightly win the absent vote and slightly lose the declaration pre-poll vote. But the biggest vote category is the postal vote, and that will be decisive. My model right now has Boele falling behind by less than 200 votes. This model assumes she polls about 44.6% of the postal votes, but would need to pick up less than 46.5% to take the lead. This remains very close.
Bullwinkel
This seat was originally a Labor vs Liberal 2CP, then that was suspended due to the high Nationals vote, but it is now back as a Labor vs Liberal 2CP with the Nationals having fallen back.
Labor is currently leading the Liberal Party by just 28 votes, with a majority of postal votes reported. The Liberal Party is only narrowly leading on those postal votes. My model expects Labor to narrowly win the absent and declaration pre-poll categories and narrowly losing the postal votes, and results in Labor winning by 118 votes. Far too close to call.
Goldstein
Teal independent MP Zoe Daniel is leading her predecessor Tim Wilson by just 95 votes. There appear to be a lot more postal votes remaining than the other categories. Depending on how many postal votes return I think she will probably end up losing by 1000-2000 votes, but she may do better on the remaining votes.
Kooyong
Currently, sitting teal independent MP Monique Ryan is leading by 992 votes.
I expect her to win the absent and pre-poll categories, and overall my model expects her to end up leading by about 1000 votes at the end.
Longman
Sitting Liberal National MP Terry Young is leading by 309 votes.
There are far fewer postal votes here than in other seats analysed so far. Labor did well on absent votes and narrowly lost the declaration pre-polls, and that’s before they gained a substantial swing this year. Overall I have Labor leading by 351 votes at the end of the count.
Menzies
Labor currently leads here by 1382 votes. Labor easily won the absent votes and narrowly won the declaration pre-poll votes in 2022, and didn’t lose the postal votes too badly. My model predicts a Labor win of just over 1000 votes.
Wills
The Greens are now 2813 votes behind Labor here. The Greens should do well on the absent votes and reasonably well on the declaration pre-poll votes, but that is cancelled out in my model by Labor winning the postal votes comfortably, leaving the final margin around the same place. This seat is close to being called for Labor.
“Top two is not clear”
Does that mean that there are potentially three or four contenders in the running?
Someone on Pollbludger mentioned that the earliest postal votes favour of the Liberals or Nationals because they are from older, conservative voters.
Is there evidence of a swing to the left (to Greens in Melb or teals and Labor elsewhere) on postals in week 2?
That’s right, it’s not clear who are going to be in the top 2. In Calwell it’s not even clear who is in the top 3.
Calwell just looks like a cluster to deal with. Monash will be a lot easier to deal with by comparison.
Having examined the booths and their primary results Abdo is leading in pretty much all of them (not by much) but the biggest hurdle is that Moore/Youhanna/Liberals are all coming second in different parts of the electorate. Moore in the Hume part of the electorate, Broadmeadows and Greenvale has the Liberals in 2nd and Youhanna in the other parts of the electorate. There’s literally a mosaic of 2PPs if it came down to sorting it out individually, which means a winner is probably going to take weeks to determine.
Whilst I wouldn’t rule out any chances that either Moore or Youhanna could get close (the Liberals have no chance), I think Labor will scrape through but the seat will be marginal.
Hi Ben, re Kooyong, if there are approximately 7500 envelopes yet to count, and the split continues 62/38 Hamer’s way, she winds up pulling about 850 votes ahead. Do you think Monique Ryan then pulls a further 1850 votes ahead from then, or have I misunderstood?
What is it with Calwell? Savage drop in ALP primary vote (obviously against the nationwide swing). Australia-wide this has been a huge ALP victory: but on the back of voter support which is broad, rather than deep. In future elections it looks like the 2PP way of predicting the results just isn’t going to work anymore.
How does the AEC deal with the more complex counts like Calwell and Monash?
Do they do 3CP/4CP counts at each booth and then work from there?
Any scope for a senate like button press where all votes are entered into a computer and then see what happens?
@Alan:
Greens votes leverage ALP votes and vice versa. Those parties stood in all seats, got 48%. Add in various Liberal and National combinations, there’s 80+% for the Majors.
Teals aren’t playing the game, if they were under Party regs, their candidates wouldn’t be able to accept $500,000 from Climate 200, no questions asked.
Get a level playing field on donations, where Liberal can accept anything too, see how the Teals fare then?
Why does this happen every election where seats are called only to be wound back by postals in the following days? The ABC and others corrects for the conservative lean of prepoll, why not for postals? Or are postals especially strong for the Liberals this tine round?
I think in the case of the teal seats in Melbourne in particular, the Liberal lean of the postal votes is even stronger than it has been in previous years (by quite a bit), so far anyway.
But in 2022 there was a pattern where the first 1 or 2 batches of postals are by far the most Liberal-leaning and they then correct back towards the normal lean. That could happen again this time too. That 65-35 break to Tim Wilson for example might start getting back to around 55-45.
Higgins was an extreme example of that last time from memory.
Much maligned polls got it right (again)
“Much maligned polls got it right” was the title of an article I wrote for the Australian Financial Review after the 1996 election, in which I said “If you massage the polls properly, they nearly always get it right”
They did it again in 2025.
The trick to getting it right is to rely upon the regression equation (derived from all elections since 1919) that links the proportion of seats won by a party to its two-party preferred vote [equation is “proportion of seats won = 2.45*TPP-0.57”].
If we leave aside the cross-bench and assume that Saturday’s election was a race for 140 seats, the ALP should have won 80 seats.
It did
QED
Don’t pretend to be an expert in this, but three possible reasons spring to mind:
1. Several recent elections have had a late swing to Labor that wasn’t reflected in early and postal votes so they have more of a conservative lean than expected.
2. Because of COVID, postal votes have had a less conservative lean than before and there is now a reversion to mean that’s not yet being picked up by the model.
3. Postal voters are older and so are not only more conservative, but more likely to be rusted on supporters of a major party, which makes them particularly rough on teals and Greens.
I think Libs will win Bradfield (56%) and Goldstein (69%) given consistent flow of postals. But here are both growing in probability.
My model still has Independent Ryan more likely in Kooyong at 61% (although reducing from original 81%) – I think the next two postal vote batches will be telling as to how this plays out. Mr Bowe is also about this.
I have no idea on Bullwinkel (exactly 50%). This will probably end up the most marginal of all seats.
I think Libs probably just retain Longman (53%).
Labor should win Menzies (78%) although there is still 9000 postals to count and would need 57.8% of remaining postals against a current flow of 57.5%. One sleeper to keep an eye on….
Adam – from my model, postals do appear to be slightly stronger relative to other votes in this election, which appears to have put some seats back in play.
Ian – I think there is a view that the first batches favour Libs more. Personally, I’m not sure that is right and it may be just down to an aberration. Postals appear to favour Libs in this seat more than others. We will know in the next two batches how that plays out. It is possible that Libs ran a better postal campaign than Ryan. Currently postals are 62% Hamer. That would need to be consistent for Hamer to win. I think it would come back a bit to say 60%. I still have Ryan as favourite but want to see the next two material batches.
Alan Calwell is one of 3 seats with high Muslim populations. Targeted by the Muslim voting groups and with less controversial candidates. Add into that the sitting mp had retired so personal vote drop off as well.
I predict libs should hold Bradfield and win Goldstein. Kooyong will probably result in a liberal win in. My opinion due to the very high number of postals still to count and the high flow rate of the 2cp of them towards the libs. Menzies should be close. Given that Menzies had a 1.1% shift after redistribution Keith Wolahan has technically got a swing towards him but he still behind not sure if he can recover enough though. Given that he’s bucked the national trend I wouldn’t rule him out recontesting. In 2028.
Bullwinkel will be lineball and I expect to under 100 votes either way. Expect a recount. If one of the candidates makes the count in calwell other then the libs they should win. Hard to see labor winning off a 30%primary
Labor has edged ahead in Bean and Fremantle. I think they’ll both be Labor retain but by a small margin. Labor could well be on the way to 90 seats at this rate.
Silver lining for the Liberals in Goldstein and Bradfield in that they’ll scrape through, just.
looks like ALP will end up in the low 90s and Lib/LNP/NAT/CLP in the mid 40s (about half of Labor)
Update as per the ABC:
Leading party by seat:
* Bean: Labor
* Bendigo: National
* Bradfield: Liberal
* Bullwinkel: Liberal
* Calwell: Labor
* Flinders: Liberal
* Franklin: Labor
* Fremantle: Labor
* Goldstein: Liberal
* Kooyong: Independent
* Longman: LNP
* Melbourne: Labor
* Menzies: Labor
* Monash: Liberal
* Ryan: Greens
Goldstein is now listed as “Liberal likely”, while everything else listed above is just “XYZ ahead” (replace XYZ with the party ahead or with independent if it’s an independent leading).
Franklin has really switched up. I thought it was done and dusted for Labor but the ABC were right to keep it in doubt as Peter George now has 48.4% of the TPP. Last time I checked Labor had about 57%, now they have 51.6%.
The Liberal lead in Bullwinkel is just 46 votes while in Bradfield it’s just 54.
@Nether Portal the ABC’s projection in Franklin appears to be an artifact of dodgy extrapolation from unrepresentative realigned booths, Kevin Bonham has been quite critical of this on BlueSky. PollBludger currently has it at >99% Labor Retain.
I’m calling Goldstein for Lib.
Called it 2 days ago
One of the theories doing the rounds about postals is that 2022 postals weren’t as conservative as usual (e.g. because those who did a postal vote because of Covid concerns were more likely to be left-leaning) and that in 2025 we’re seeing something of a reversion to normal.
Separately, it wouldn’t be a huge surprise to see unusual behaviour of postal (and absentee) votes in Queensland and the NT because the election was on a long weekend there.
This seems to be a Victorian thing more than anything BT, as postals in late counting for states in other seats don’t seem to be markedly more conservative than 2022. Perhaps that’s due to the nature of the extended COVID period for Victoria.
I would’ve thought that there are more postal voters and absentee voters this election due to back to back long weekends and thus more people going on holidays. If this is true then the average voting age of non-ordinary voters wouldn’t be as high as pre-pandemic levels.
The difference in vote type is like night and day in Goldstein. 52/48 Zoe Daniel’s way on election day (and hence the election night partying), but 36/64 on postals. Antony Green said that the Liberals’ postal vote campaign was one thing they got right.
Independent just in front in Bean tonight – a large dump of pre-polls from Tuggeranong generated a preference flow that put her back in front
@Tommo9: You can’t rule out a shock Labor defeat if either Moore of Yohana (especially Yohana) finishes in the final two. Both independents Carly Moore and Joseph Yohana have a strong chance to pass the Liberal Party and finish second during the distribution of preferences.
If the Liberal Party finishes second Labor will win easily. If Carly Moore finishes second, Labor will likely win narrowly because the Greens recommended preferences to Labor ahead of Moore on their HTV cards while the Liberal Party, One Nation and Yohana recommended preferences to Moore ahead of Labor.
I think Yohana would be the biggest threat to Labor if he finishes in the final two. The Greens HTV ranks Yohana higher than Labor, Liberal and Moore and the One Nation HTV ranks Yohana higher than Labor, Liberal and Moore, Greens and Moslih. Moore’s HTV put Yohana second. Therefore, there’s a possibility that during the distribution of preferences, Greens and One Nation preferences will push Yohana ahead of Moore, and Moore’s preferences will then push Yohana ahead of the Liberal Party and get Yohana into 2CP. Liberal preferences will flow strongly to Yohana threatening Labor’s hold on the seat.
On current primary vote figures, Yohana needs 66.4% of preferences to win should he make it into the top two. 66.4% of preferences should be fairly easy to achieve given that Greens, Liberal, One Nation and Moore have both put him ahead of Labor on their HTV cards. Independent Andrew Wilkie got 69.75% of preferences when he defeated Labor in Denison in 2010.
Note that although the rate of minor party and independent voters following HTV cards is low, independent voters tend to preference another independent over any party. Many Greens voters may still preference Labor second despite Greens’ HTV recommending them to preference two independents above Labor. Moslih recommended preferences to the Greens ahead of any other independent on his HTV, however when he’s excluded his preferences may end up favouring Yohana because Yohana and Moslih are both independents from diverse backgrounds, offering another pathway for Yohana to finish ahead of Moore during the distribution of preferences.
It would cause a big political earthquake for Labor to be defeated in Calwell, used to be one of the safest Labor seats in Australia, by a fourth-placed independent with a primary vote of just 11.8%. However, this federal election has no shortage of shock results, in particular Adam Bandt’s defeat in Melbourne. Maybe Calwell will just be another one.
Independent how-to-votes aren’t going to receive high follow rates. The Liberal party’s how-to-vote matters a lot in these contests but the Greens matters considerably less because the follow rate is under 10% as opposed to 30-40% for Liberals. Independents will be likely even lower because they have much less material that reaches voters, especially those voting outside of a booth. Of course this can’t be called because of the difficulty of figuring out who reaches the final two and how the votes flow to them but once it’s figured out I doubt this will be particularly close.
I would certainly question the assumption that independents will preference another independent over Labor, particularly to the point of 66% preference flows. Some people will be the “put the majors last” kind but they are far from a majority and quite a lot will simply not put relatively unknown independents over a party whose name they recognise.
@Adda There is evidence that suggests independent voters tend to preference other independents over any party. What’s more, when the Greens recommend preferences to independents over Labor, Greens preferences also tend to favour independents over Labor.
For example, in the 2022 federal election in Groom, when independent Kirstie Smolenski was excluded and Labor, LNP, One Nation and independent Suzie Holt were left, 64.42% of Smolenski’s preferences flew to fellow independent Suzie Holt despite Smolenski having issued an open ticket. Holt got 72.6% of preferences despite polling only 8.26% of primary vote. When the Greens were excluded, Holt and Smolenski received the largest share of Greens preferences, just like the Greens HTV put the two independents ahead of any non-Greens candidates. In the 2022 federal election in Bradfield, when independent Janine Kitson was excluded and Labor, Liberal, Greens, UAP and independent Nicolette Boele were left, 59.53% of Kitson’s preferences flew to Boele.
In Calwell where there’s massive disillusionment with both major parties, most voters switching to minor parties and independents were disillusioned with both major parties and therefore were likely to push the major parties down their ballot. It’s reasonable to assume that many would have preferenced some independents or minor parties, including those they have not heard of, before preferencing major parties.