Victoria Aggregate Labor 22.2 Coalition 31.2 Green 15.0 One Nation 21.6 Others 10.0
2PP estimate 54.4 to Coalition, shadow-2PP 50.5 to One Nation vs ALP (subject to Coalition recommending preferences to One Nation)
Currently Coalition would most likely win a majority, perhaps with around 50 seats
-----------------------------------------------------------------------------------------------------------------------
I last had time (sigh!) to write about the Victorian election polling outlook way back in July. At that point the Coalition was the better placed of the two major parties but were still a few points off a high enough primary vote for majority government.
What has happened since then in polling overall is that Labor's leadership change from Jacinta Allan to Ben Carroll has delivered less than nothing for the government. In fact Labor's polled primary vote share is down, One Nation's is also down, and both the Coalition and the Greens are up. None of these changes are massive, but they don't need to be - as the Coalition pushes into the low 30s its chances of majority government become serious, especially if One Nation goes down. In recent months the Coalition has gradually edged from being in the box seat but not yet up to a majority in live polling, to a majority being more seriously likely, and now there is even some focus on the potential for a blowout and for Labor's vote to be devoured on all sides. Ben Carroll has continued to make the increasingly ludicrous claim that polling shows the Coalition cannot govern alone, even while polling suggests that if the election were held right now it almost certainly would. No doubt at least three people in some focus groups somewhere have implied this scare campaign is solid but I wonder if what it's really doing is scaring voters who actually care to the Coalition as the only party that looks like it might get a majority.
Steve Bracks this weekend made the even more ludicrous claim that the three main parties are all polling around 30% - neither Labor nor One Nation has broken 25 in any poll since August! Labor's high for the year in any poll is 28 and One Nation's 27, and both of those were a long time ago. Bracks also claimed this was the first genuine three-cornered contest in Victorian electoral history, which would be news to those who voted in pretty much any Victorian election between 1935 and 1955, or watched the formation of government in that time.
In the polling in the last few months Victorian Labor has overall looked more and more like an ex-government, and there is a sense of chaos similar to the final days of Rudd 2 as the party tries to work out how you sell yourself after a messy leadership split when the new leader has not had time to get settled, nor the party to settle its campaign around him. Much of their output seems to be chasing voters lost to One Nation with weird law and order policies that for anyone even listening are likely to be seen as too little too late. That said, not all of the readings are bad and there have been some polls (especially the Resolves) where one could argue that if everything went well the government might survive on those numbers. Overall though things have been getting worse, and it is historically not surprising. A third term federally dragged state government, with the federal government's own polling increasingly iffy (no surprise there given high petrol prices, an interest rates rise and recession talk) and with the state government having made its second leadership change of the term only a few months out. Ben Carroll is already polling very unflattering personal ratings with a Newspoll net satisfaction of -28; not as bad as Jacinta Allan's final -46, but still no precedent for survival.
This is what my estimates of the implied Labor vs Coalition 2PP have looked like for the government in the last few months. This may tighten again in the final weeks as the Coalition comes under more serious scrutiny as the government-in-waiting but a recovery from here for Labor to form any kind of government would be remarkable. Nonetheless if Labor do pick up a few points and the polls have them too low for some reason, it can't be totally ruled out.
At the head of this post I have some aggregated estimates of polling, using basically the same methods as my federal aggregate except for the following variations. Firstly I have set the decay period (in which a poll loses about 40% of its weighting) to 10 days instead of 7 because of the relative scarcity of polling (Labor's numbers would be even worse if I had not done this); I will reduce it to 5 for the final week. Secondly while SMS polls are banned from my federal aggregate, here I am using them but downweighted by 50% and with a maximum of one per pollster. Thirdly I am using an arbitrary 55% Others to Labor flow.
Current Exclusion
I am currently not aggregating the Lighthouse Political Monitor tracking poll reported by the Herald-Sun and conducted by News Corp's Growth Intelligence Centre. The reason for this is that the Herald-Sun has persistently displayed very poor poll reporting transparency and a lack of standards about what it calls a poll, and there is not yet sufficient public information to establish firmly that this is a scientific poll. In all probability it is, but I can't be sure. I will however test it in my post-election poll ratings, but for me to include it in an aggregate I need to at least know what polling method it employs (online, phone, etc) and to have some indication of the question framing. I am also concerned that the Herald-Sun continues to allow its poll reporting to be done by James Campbell. While he is not alone in being terrible at it (and often in precisely this manner, ie not enough basic methods details) he has what should in any sane business be a dealbreaking conflict of interest as the husband of a Liberal candidate for the election.
There is a lot of uncertainty about what will happen with preferences from both One Nation and others at this election. Respondent preference samples have at times found One Nation flowing to the Coalition as strongly as the high 80s (but a cautionary note that some low-grade polls had the same thing at the federal election and it turned out not to be the case). On the other hand, the right wing of the Others vote could be so hollowed out by competition from One Nation, and the Others vote so dragged left by things like Victorian Socialists running everywhere, that the flow from Others to Labor might climb to something like 70%, or even more.
There was an interesting possible pointer to the latter in the most recent Newspoll, where a seemingly remarkable 77% of Others voters chose "Labor majority" or "Labor-Green" over "Liberals-Nationals-One Nation" or "Liberal-National majority" when asked who they preferred to govern. Perhaps however the framing was a little harsh on the right as the option including One Nation could be read as strictly a formal coalition option that excluded an L-NP minority government supported by One Nation. That at least might partly explain why a third of those with a Coalition majority as their first choice next preferred one of the Labor options over the Coalition-One Nation option.
For what it's worth though, only the Redbridge/Accent respondent preference polls are lately showing large differences to my last-election-ish estimates off the primary votes. The last two Redbridges came out with 56-44 two-party leads to the Coalition off numbers that my estimates put around 53.4. Other respondent polls have been landing within 1% of what I have.
Overall Picture
My model off 54.4% 2PP for the Coalition (Labor 22.2 Coalition 31.2 One Nation 21.6 Greens 15.0) puts them on about 53.7 2PP seat wins. This model was rerun after today's DemosAU reading with Labor on an 18% primary and I expect Resolve shortly which may be more moderate. The Coalition winning that many seats on 2PP doesn't mean they would necessarily win all those seats. In some outer suburban and regional fringe seats, it's quite possible the Coalition would win the 2PP in seats that were actually won by One Nation. But all else equal for the Coalition to not win a majority, One Nation would have to achieve this in a lot of those seats. Considering, for instance, the seats that are at least considered to lean One Nation in William Bowe's model, my expected Coalition 2PP-winning 2PP total is currently 10.3 seats. So One Nation would have to win nearly all of these, or win some seats in which it is an underdog, to stop the Coalition getting a majority.
There is also an emerging possibility that even with a vote over 20 One Nation may not win more than a few seats. Labor going down and the Coalition going up on primaries is bad news for One Nation. To win seats they firstly need to be in the top two after preferences, but they also need either the Coalition to be third or to have a lead over the Coalition that will hold off Labor preferences. And with the Coalition polling well and Labor poorly, and One Nation slipping down a bit, it's possible that those stars just won't align in any serious number of seats. For sure if One Nation are getting over 20 statewide they will be breaking into the high 30s in some seats. But if those seats are deep rurals where Labor is third or fourth and the Nationals are also around the same vote, One Nation doesn't generally win those seats. Based on SA, however, it shouldn't be ruled out in vacant rural seats.
This I think is what Kos Samaras is getting at when he has One Nation getting just a single seat (Morwell) to the Coalition's 52 off quite similar primaries to my aggregate. That could well be understating it because (i) Redbridge/Accent has very weak respondent flows to One Nation off the Coalition and will tend not to flag them as winners in any seats where the 3CP is Labor leads One Nation leads Coalition (ii) I would expect natural variability in the results from any projections such that One Nation randomly gets luckier than expected in a few seats. Nonetheless there is an argument there.
There is a common belief that Victorian Coalition voters are unusually moderate and hence very averse to preferencing One Nation. Senate preference flows provide no evidence this is the case. In 2025, of those Victorian Senate voters (above and below the line) indicating a preference between Labor and One Nation, 73.8% preferred One Nation. This is lower than NSW (78.2%) and Queensland (77.1%) but the same as Western Australia, and higher than South Australia (71.2%), Northern Territory (70.3%) and Tasmania (63.4%). We've already seen that Coalition voters in South Australia did quite readily give preferences to One Nation in the state election. In cases where the Coalition finished third, 66.6% of Liberal exclusion preferences flowed to One Nation over Labor there, and this is a flow that includes votes the Liberals picked up from other sources, so the flow from 1 Liberal votes would have been stronger.
It's true that Alex Antic and co have destroyed the Liberal brand in South Australia driving away many moderate voters, but that was also true at the federal election. It's also the case that Labor in the SA state election were a far less offensive proposition for traditional Coalition voters than they are in Victoria. So I don't think there's the slightest thing in this argument ... unless the Coalition decide not to recommend preferences to One Nation. Predictably the federal seats where the flow of Coalition preferences to One Nation over Labor was weaker in the federal election were generally seats where Labor polled strongly and/or One Nation weakly. Calwell was one exception but there a lot of Labor's support was taken by a very strong vote for Australia's Voice most of which returned to Labor as preferences. In the Goldsteins and Kooyongs the flow was normal, but in seats of that kind One Nation will be eliminated anyway.
I have not had the time yet to attempt a detailed Lower House seat model of One Nation prospects in Victoria; it is more difficult for me than it was in SA because One Nation polled so negligibly in the Victorian upper house in 2022. Between various predictive enterprises out there (some are nowcasts, some are forecasts) it's interesting to look at the variations in where One Nation is considered favourite.
(At the time of writing Australian Election Forecasts doesn't have One Nation as a forecast favourite in any seat but still has a median total of seven.)
The table above compares the Pollbludger model, Australian Election Forecasts nowcast, Kos Samaras' recent AFR assessment of the current polls, Swing Is On, Earth Elects and current betting. For the Pollbludger column S equals Solid and L equals Lean. For everyone else O indicates an odds-on favourite (after adjusting for bookie rake for the betting) and F is any kind of other favourite or favourite of unspecified probability. In all, 24 seats have come up in at least one of these sources, but only the usual suspects of Bass, Melton, Morwell and Ripon in half or more of them.
No comments:
Post a Comment
The comment system is unreliable. If you cannot submit comments at all 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. Emailed 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 (sometimes I do not see notifications for an incoming comment). Comments never appear immediately as I have to manually approve comments prior to publication.