RedBridge Group/Accent Research have a large-sample Victorian state poll in the Financial Review, showing One Nation overtaking the Coalition since the last such poll in February. The primary votes all but record a three-way tie between Labor, up a point to 26%, One Nation, up three to 27%, and the Coalition, down two to 26%, with the Greens steady on 13%. Two-party results have the Coalition leading Labor 54-46, and Labor leading One Nation 52-48. Jacinta Allan records 17% favourable, 24% neutral and 59% negative (which the Financial Review says are the worst recorded for any leader other than Anna Bligh and John Cain, though on what basis is unclear), and Jess Wilson is on 31% favourable, 51% neutral and 18% unfavourable. Pauline Hanson was thrown in for good measure, recording 37% favourable, 17% neutral and 46% negative. The poll was conducted June 17 to 28 from a sample of 5516.
Further electorally relevant news from Victoria:
• The Liberal Party state executive meets tomorrow to decide whether to remove Moira Deeming as lead upper house candidate for Western Metropolitan. The Age reports a “broad consensus” within the party that she should be dropped, as there would have to be to get the required agreement of 75% of the executive’s roughly 20 members. Deeming’s defeat in an April preselection vote by restaurateur and Indian community figurehead Dinesh Gourisetty prompted a backlash on the right, and she was shortly able to recover the position when Gourisetty was obliged to withdraw. However, last week’s events appear to have damaged Deeming’s prestige. Pauline Hanson bluntly told Melbourne station 3AW today that the party “don’t want her”, and hitherto strong supporter Peta Credlin, who is married to state executive chair Brian Loughnane, “has let it be known that she is now done with Deeming”, according to Chip Le Grand of The Age.
• Sumeyya Ilanbey of the Financial Review reports Labor has written off nine seats, including several where the threat would presumably be from One Nation: Sunbury, Ashwood, Melton, Yan Yean, Glen Waverley, Hastings, Pakenham, Bayswater and Bass. The party is confident of 19 seats and targeting another 26, suggesting it is plotting a bare minimum path to a majority of 45 seats out of 88.
• Kieran Rooney of The Age reported last week that Labor internal research indicated the Nationals seats of Gippsland South, Murray Plains and Gippsland East were likely One Nation wins, with Shepparton, Benambra and Lowan “leaning orange”. Also in the latter category were Berwick, the Melbourne fringe seat held by former Liberal leader Brad Battin, along with Liberal-held Narracan, Warrandyte, Eildon and Rowville. No Labor seats were identified, but One Nation was acknowledged to be “a major problem for both sides of politics”. Liberal sources expressed high hopes for gaining Niddrie and Macedon from Labor.
• Ryan Bourke of the Herald Sun reports One Nation “remains without a frontrunner” after an apparent failure to recruit as candidates Adam Giles, former Northern Territory Chief Minister and now chief executive of Gina Rinehart’s agricultural arm, and Heston Russell, former commando and founder of a party that ran at the last state election as Angry Victorians.
My estimates of how the Victorian Upper House might go on those numbers.
Northern Metropolitan: 2 Lab, 1 Lib, 1 ON, 1 GRN
North-Eastern Metropolitan: 1 Lab, 2 Lib, 1 ON, 1 GRN
Western Metropolitan: 2 Lab, 1 Lib, 2 ON
South-Eastern Metropolitan: 2 Lab, 1 Lib, 2 ON
Southern Metropolitan: 1 Lab, 2 Lib, 1, ON, 1 GRN
Northern Victoria: 1 Lab, 1 Coalition, 3 ON
Eastern Victoria: 1 Lab, 2 Coalition, 2 ON
Western Victoria: 2 Lab, 1 Coalition, 2 ON
So totals,
Labor: 12
Coalition: 11
One Nation: 14
Green: 3
This is off-the-cuff since a quota in the upper house is 16.7%. And with Group Ticket Voting still in place, it can be assumed that the Coalition and One Nation will effectively preference each other just as much as Labor does with the Greens.
Of course there’s probably going to be up to 7 microparties also elected out of these numbers at the expense of the above four from like 0.5% of the vote as they continue to play the Glenn Druery musical chairs.
But overall, there’s going to be a solid Coalition-One Nation majority in the Upper House from these numbers.
If only there was something Labor could have done in the last twelve years to prevent the right’s GTV-assisted takeoever of the upper house from happening.
If only.
I wonder with the current polling if Jess Wilson and Victorian Liberals could move their campaign and election promises further to the right to appease One Nation People to switch to vote Liberal to consolidate the conservative vote (or at least preference them ahead of Labor)?
Trying to ‘appease’ One Nation would likely backfire (consider Angus Taylor’s current dire position) and could jeopardise the strong position the Victorian Liberals have in two of the five metropolitan regions (according to Redbridge). Jess Wilson would be an unlikely candidate to endorse a Coalition campaign veering to the right, considering both her faction and her seat (Kew).
In their June poll, DemosAU identified the two biggest issues in Victoria as ‘cost of living’ (33%) and ‘crime and violence’ (29%). Campaigning hard on these would likely be more successful (for any of the parties).
I’ve used that Redbridge breakdown of the lower house vote grouped by LC region, to calculate the swings at an LC level and apply that to each seat within the LC region.
Obviously there are still a lot of caveats here and it’s still dubious for these reasons:
1. While not uniform at state level, it’s still uniform across 11 seats that may still have quite different geographic variances (especially in Northern & Western Victoria, for example, where the ALP base is so low in rural seats that the swings will be more concentrated in regional centres);
2. Doesn’t factor in things like local issues or incumbency;
3. It doesn’t work well in seats where an IND had a significant vote.
Also just to explain the methodology, because ONP barely ran in 2022 but VIC had a whole lot of right-wing minors, I calculated the ONP swing based on comparing their polling to the combined right-wing minor vote from last time, since it’s very likely to consolidate behind them this time.
Here’s the seat breakdown I get:
ALP 34
GRN 4
LNP 26
ONP 13
Tossups – 11
Of the 11 tossups, 9 of them are ALP v either LNP or ONP, 1 is ALP v GRN and 1 is LNP v ONP.
Of those 9 ALP vs LNP/ONP ones, there are a handful where the decisive race is actually between LNP & ONP at the 3PP stage, and ALP would go on to retain vs ONP but lose vs LNP.
So on these numbers, Labor cannot form majority government, 44 would be “ceiling” even if they won all 10 “tossups” they are in which is very unlikely.
Overall it’s basically a 38 Left (ALP/GRN) vs 39 Right (LNP/ONP) starting point with 11 tossups to decide the balance. Which is pretty close to where I would expect it to be.
Another caveat: there are a handful of seats where I completely disagree with what the numbers gave me, but I didn’t change them. One example is that it has ALP retaining Northcote, a couple of others I mention below.
The LC region that surprised me the most was North-East Metro, because this poll shows a -15 primary vote swing for the ALP there which is the biggest swing anywhere. North-East Metro is absolutely not where I expected the biggest swings to be.
This has even made Eltham & Bundoora a tossup, which I disagree with. I have left them as such though but I think they are 2 of the ‘Tossups’ I would pretty comfortably give to the ALP.
The other big finding is due to ALP faring better in 2CPs vs ONP than LIB, ONP surpassing LNP helps Labor retain quite a few seats.
I won’t go through every seat, but at an LC level here’s a summary:
Southern Metro – Not much change (Ashwood becomes a tossup, Greens regain Prahran)
Northern Metro – The only change is ON narrowly gaining Greenvale. But as mentioned above I disagree that Labor will retain Northcote. I’d personally make that a GRN Gain (but haven’t).
Western Metro – Lots of movement. One Nation gain two, Werribee becomes an ALP v ON tossup, while 3 seats (Niddrie, Sunbury & Sydenham) have ONP & LNP tied at the 3CP, where Labor would retain vs One Nation but lose to the Libs. So it could range anywhere from ALP retaining 9 of 11 seats, to ON & LIB gaining 3 each while the Greens snatch Footscray.
North-East Metro – This was the surprise where Labor could potentially be wiped out and lose 6 of the 8 seats they hold (3 tossups & 3 LNP gains among the 6 seats Labor won in 2022). However, I do disagree with a couple of these results, in particular Bundoora & Eltham.
South-East Metro – Labor hold up surprisingly well here, only losing 1 seat: One Nation narrowly gain Cranbourne.
Western VIC – Another surprise with Labor holding up ok, but One Nation gaining the 2 obvious ones (Melton & Ripon, in thumping fashion).
Northern VIC – One Nation gain 3 (2 off LNP, 1 off Labor) while 1 becomes an LNP v ONP tossup (Mildura). Jacinta Allan retains Bendigo East on the numbers. But, this is a region where I think the uniform LC level swing won’t apply, because Labor’s starting base is so low in some of the rural seats that I think the swing will be more concentrated to the regional centres. Therefore I think Bendigo East & Macedon in particular are two seats where I don’t agree with the numbers: they will be tossups.
Eastern VIC – One Nation clean up here with 4 gains, while Hastings is an interesting result where it’s hard to tell if the 2CP will be ALP-ONP or ALP-LIB. It’s a clear LIB gain if it’s the latter but a tossup if it’s ALP-ONP. I don’t agree with that though for the same reasons mentioned above in Northern VIC, seats like Hastings are probably where the swing against Labor will be more concentrated.
So there’s the summary I found. Like I mention a few times, they aren’t my predictions as I emphatically disagree with a few for the reasons I outline above. That said, it doesn’t change the overall outcome much because the seats where I think the numbers are too favourable to Labor (Bendigo East, Macedon, Hastings) are offset by the ones where I think Labor will do better than the numbers say (Bundoora, Eltham, Box Hill).
So I think around 38-39 seats each for the ALP/GRN bloc and the LNP/ON bloc, with around 10-11 tossups, is probably pretty accurate. And Labor will probably want to face ON instead of LIB in the 2CP of as many outer suburban and regional centre contests as possible, it could save them a good 4-5 seats.
Any chance ALP vs GRN contests flip if Libs preference ALP or ALP primary too cooked to matter?
I think the ALP v GRN seats are the ones that I really don’t think my above model of applying those uniform LC-level swings to all 11 seats within an LC works for.
Firstly, I think they are the seats that will deviate most from the uniform swing across the LC. If you look at Western Metro for example, Footscray is the very obvious “odd man out” in a region full of seats like Sunbury, Sydenham, Point Cook, Tarneit, Werribee, etc. So you end up applying a huge ON swing (due to the outer seats) in an inner seat that is very unlikely to have one.
And secondly to your point, how the LIBs will direct preferences is still an unknown, and also how ON preferences flow between ALP & GRN is pretty unknown too, so the preference flows for the ALP v GRN seats are also the least informed guess.
Whereas the GRN v LIB seat (Prahran) is far easier: ON preferences will strongly flow to LIB while ALP preferences will strongly flow to the Greens. It’s much less clear cut in the ALP v GRN seats.
Preston & Pascoe Vale were a mess at the 3PP stage and further complicated by INDs doing well in 2022, but the model gave Labor comfortable retains vs any 2CP opponent.
The problem in Northern & Western Metro was that there were huge ON swings, even when compared to the overall right-wing minor vote in 2022, of +20 in western and +15 in northern, which would have been sourced more from the outer seats but I had to apply them to Melbourne, Brunswick, Richmond, Northcote & Footscray where they will absolutely not occur to that degree (which also means the ON swing is probably underestimated in the outer seats).
This made them really hard to work out so I really don’t put any faith in those results. Although Melbourne, Brunswick & Richmond at least had the outcome I would expect (Greens retains) whereas it really just left Northcote & Footscray as a mystery.
I wonder if One Nation choose Bernie Finn as their leader for Victoria , I wonder how it would fare?
He’d narrow their appeal back to their ultra-right base very quickly.
Trent, that’s certainly a fascinating analysis and highlights how misleading the poll’s overall ALP-L/NP 2pp could be. Agree that applying uniform swings across each of the eight regions is problematic, given all available polling indicates the One Nation swing increases from the inner suburbs to the outer suburbs to regional areas. Add to that how different voter behaviour is shaping for this election.
Eastern Victoria is the most surprising here – given the One Nation (32%) and Coalition (28%) primaries are relatively close here, with Labor far behind in third (22%), I would expect the Coalition to romp home on Labor preferences here, rather than One Nation taking four seats.
The corresponding article in the AFR also has ALP-L/NP 2pp votes for each region, some of which are obviously more relevant than others as ALP-L/NP contests:
– N Metro – ALP 61-39
– W Metro – 50-50
– W Vic – 50-50
– N-E Metro – L/NP 59-41
– S Metro – L/NP 59-41
– S-E Metro – L/NP 52-48
– N Vic – L/NP 63-37
– E Vic – L/NP 60-40
Substituting One Nation for the Coalition would surely reduce the 2pp margin against Labor, but suggests seats in W Metro, W Vic and S-E Metro will be very difficult to predict.
And on raw primaries, S-E Metro looks the most complicated give how close all three parties are on primaries.
As for preferences, I would expect preference flows between the Coalition and One Nation to be significantly stronger than in SA, given the relative popularity of the two incumbent state Labor governments. Nepean might give a better indication of One Nation preference flows to the Coalition in Victoria, though this contest was slightly different given the final count was not actually an ALP-L/NP contest. One Nation is also actively campaigning against Labor in Victoria (rather than against both major parties, as in SA) and the Coalition seem to be signalling they will advise preferences to One Nation ahead of Labor (at least in most seats).
It will be interesting to see the next set of Victorian polls. I think Resolve and Roy Morgan should both report this month. The next DemosAU poll would also be interesting, given it was the Coalition’s strongest in June.
To clarify on Eastern Vic, those One Nation gains include gains off Labor, not just the Coalition. And one of the gains off the Coalition (Narracan) comes with a significant asterisk because there was no Labor candidate for the 2022 result, and therefore no Labor preferences to distribute in this model.
Most of the model’s results for Coalition-held Eastern VIC seats support your theory that Labor preferences help the Coalition retain.
More on that: in Eastern VIC it had the Coalition comfortably retaining Nepean, Gippsland South, Gippsland East, Evelyn and Mornington, and the Coalition still leading One Nation on primary votes even before the Labor preferences are distributed in all of them too.
Three of One Nation’s gains were ALP v ONP contests where it was Coalition preferences distributed rather than Labor ones: Bass & Pakenham (obvious gains off Labor), and Morwell which the Coalition lose because they drop to third place, but if they remained in the 2CP they would likely retain it.
Narracan is an outlier that I put an asterisk against because I only had the 2023 special election result to go by (unlike Werribee & Prahran where I could use the 2022 result), and there were two key factors:
1. No Labor candidate, so no Labor preference distribution;
2. More importantly, a combined right-wing minor vote over 25% which the ON swing gets applied to.
This meant ON ended up with a huge primary vote lead over the Coalition even before Labor preferences would have been considered; but it’s really an asterisk because in a general election with a Labor candidate on the ballot, the dynamic would be very different.
@No Favour, also on preference flows, I generally used the following:
80-20 split of ON preferences to LIB in ALP v LIB seats
60-40 split of LIB preferences to ON in ALP v ON seats
My reasoning for the first one is, like you say, ON preferences will probably flow stronger in VIC than ON due to ON campaigning more just against Labor and to change the government.
For the second one, my reasoning for a weaker preference flow is firstly because I think that in the midst of a ON surge, the remaining LIB voters will probably be overrepresented by:
a) More moderate, classic Liberals who are repelled by One Nation;
b) Swinging centrist voters who have swung from ALP to LIB between 2022 and 2026
Whereas the more conservative or reactionary Liberal voter is the type whose primary vote is most likely to have already jumped from LIB to ON.
Secondly, the fact that the Redbridge poll had the LIB 2PP a whole 6 points higher than the ON 2PP supports the above assumption. And I found that in the model I used, those preference flows worked pretty well. In seats where I calculated both the ALP v LIB and ALP v ON 2PPs, I found there was usually anywhere between a 2-5 point difference between the two, so less than Redbridge’s 6 point difference but still factoring in a stronger flow one way than the other in the same direction as the Redbridge one.
A few good examples where the 2CP matchup impacts the result:
Sydenham had a 51-49 ALP Retain vs ON but a 48-52 ALP loss vs LIB
Niddrie had a 53-47 ALP Retain vs ON but a 48-52 ALP loss vs LIB
Sunbury had a 52-48 ALP Retain vs ON but a 47-53 ALP loss vs LIB
Kororoit had a 52-48 ALP Retain vs ON but was a 50-50 tossup vs LIB
The same pattern was in seats where the result wouldn’t have differed. For example it has ALP retaining Frankston 54-46 vs ONP but only 51-49 vs LIB.
Agree – Coalition preferences to One Nation will likely be significantly weaker than One Nation to Coalition. But in South Australia, SA Liberal preferences flowed about 2-1 to One Nation against Labor – I would think Coalition preferences would flow at least this strongly in Victoria?
The 2pp votes in W Vic and W Metro must be weaker for One Nation against Labor than the poll’s reported 50-50 ALP-L/NP and suggests Labor could successfully sandbag seats here if the Coalition stays pushed to third place.
Makes sense for Eastern Victoria. One Nation’s best chances in the regions would be in regional seats where the Labor vote holds up against the Coalition. The Coalition would be frustrated if Labor preferences only protect their existing seats here.
Also – Labor seem to be holding up surprisingly well in their seats in Southern Metropolitan in your model? For example, Ashwood and Bentleigh have Labor 2pp margins of under 10%, based on primaries of 40% and 46% respectively, which Labor will struggle to get even close to if their S Metro primary is on average only 25%. Additionally, Ashwood is one of the nine seats Labor have reportedly “written off” based on their own polling.
In Southern Metro, the “uniform” primary vote swing across those 11 seats for Labor was only -6 (Labor got 31% across those 11 combined seats in 2022), so they are holding up there relatively well compared to their statewide primary vote swing sitting in double digits.
Labor also only hold 4 of the 11 seats in Southern Metro and two are on double digits, so they didn’t have much to lose and I got this:
Bentleigh 52-48 retain
Ashwood 50-50 tossup
Albert Park 55-45 retain
Oakleigh 56-44 retain
Honestly they all seem pretty realistic to me. I did Southern Metro first and thought wow, this model looks pretty accurate. Then I did Western Metro and thought WTF are some of these results?? Haha.
Interestingly in Southern Metro, the primary vote swing (-6) and 2PP swings were very similar to each other, pretty much all the ALP v LIB seats had 2PP swings between 6-7%. Whereas in most other regions, the 2PP swing was quite a bit less than the primary vote swing. But I’d expect that in Southern Metro which has the highest Lib vote and lowest One Nation vote, minimising preference leakage.
Hawthorn and Kew were exceptions because they became clear LIB v IND contests, and the Libs got 2CP swings of about +2 in them.
Malvern, Brighton & Sandringham all revert to having double-digit LIB margins again (all in the 60-63 range) based on these results too.
Trent – That also makes sense. If Labor’s small S Metro primary vote is concentrated in their existing seats they may be able to do well in them. Though clearly they are not optimistic for Ashwood.
Asha – The Age is reporting Victorian Labor are planning to finally abolish group voting tickets. I imagine this will sail through parliament before the election with the support of the Coalition and the Greens.