Poll Roundup: The Long Slow Slide Continues

2PP Aggregate (Last-Election Preferences) 50.6 to Coalition (+0.3 this year)
With One Nation Adjustment 51.1 to Coalition

If polls are accurate, either side could win election "held now", most likely well into minority

We've had quite a few pollsters out early with 2025 being an election year and the nine federal polls so far provide plenty to talk about for a roundup piece.  There have been four weekly Morgans plus one each from Newspoll, YouGov, Freshwater, Resolve and Essential.  All nine have had the Coalition ahead on their headline 2PP measure, and six have had the Coalition ahead on my estimate of respondent preferences.  Nonetheless my own estimate still has the last-election 2PP rather close.  

Voting Intention Fine Details

This section has a little more wonky detail than usual, mainly to explain the ins and outs of why my aggregate is better for Labor than the headline preferences in the current batch of polls, but also to explain why it could easily be a bit worse.  While I have Labor on 49.4 by last-election preferences, it could arguably be 49.2.  

Morgan opened with a 53-47 Coalition lead by respondent preferences but the primaries (ALP 31 L-NP 40.5 Green 12 IND 9.5 ON 3.5 other 3.5) came out at 50-50 by last-election preferences; the cause was a rogue Greens voter preference sample.  This was the largest lead to Coalition in any headline 2PP except for one March 2024 Essential which was a similar case of volatile respondent preferences.  After a more normal second week poll from Morgan (51.5 to Coalition by respondent prefs, 50.5 by last-election) things turned more clearly bad for Labor with a 52 by both methods to Coalition in the third week, followed by the current 52 by respondent prefs (51 last-election) off primaries of ALP 29.5 L-NP 40.5 Greens 11.5 (lowest since Dec 2023), IND 9 ON 6 others 3.5.  Morgan has the Independent option on the readout everywhere, which drives it up compared with what would be expected at an election.  The third-week primaries (28.5-42) tied the mid-December Morgan for the largest major party primary vote gap from anyone.

The first non-Morgan poll out was YouGov on 17 Jan with a 51-49 to Coalition off primaries of ALP 32 L-NP 39 Green 12 ON 7 others 10.  While this sounds bad for Labor my last-election estimate for these primaries comes out as 50.2 to Coalition so my aggregate didn't hit them as hard for it as those that use the headline 2PPs.

This was followed by Freshwater, a poll that has usually been better for the Coalition than other polls taken around the same time as it, but in the last few that trend has stopped being clear.  Freshwater very briefly caused my aggregate to jump to a 50.4-49.6 lead to Labor, which might seem really odd given the poll was 51-49 to Coalition (ALP 32 L-NP 40 Green 13 others 15) but this is why it happened.  Firstly, those primaries would normally imply a 50-50 last-election 2PP (I get 50.1 to ALP).  While there can be explanations for this sort of difference, including unusual breakdowns of the Others vote combined with rounding, I wasn't able to find any recent public statement on how Freshwater calculates its preferences, so my aggregate calculates its own.  (I believe they were using last-election but am unsure if this is still the case - even if so there are a range of different approaches to converting Independent responses to preferences.)  Secondly as Freshwater had been good for the Coalition recently, my aggregate then interpreted a benign 2PP estimate for this poll as good for Labor, and there wasn't much data in the aggregate at the time.  

Although the shift in how Freshwater has compared to the aggregate in its last three polls was significant enough to take a shorter-term view on that relationship, the shorter-term view my aggregate allows is the most recent six polls from the pollster.  And because some of the Freshwater polls in spring 2024 were really bad for Labor, using the last six to set Freshwater's apparent house effect relative to other polls doesn't actually change it very much.  My aggregate's unusual response to this one poll mostly washed out as more polls came in, but without that the current reading would be 50.8 to Coalition.

Essential was hailed as a good poll for Labor by those on social media wishing to discount the other polls, but they only had eyes for Albanese's net zero personal rating in it and didn't even notice that the headline 2PP+ had the Coalition ahead 48-47 (=50.5-49.5).  Essential's primary votes are deflated compared to other polls by leaving "undecided" in the mix - the Coalition's 37 (effectively more like 39) was their highest from Essential of the term.  The rest: Labor 30 Greens 12 ON 7 Ind/Other 7 the putative UAP (subject to High Court case re its name) 2 and undecided 5.

Resolve was 51-49 by last-election prefs but 52-48 by respondent off primaries of ALP 27 Coalition 38 Greens 13 ON 7 IND 10 (on readout everywhere plus forced choice, so likely to be inflated compared to an election scenario) and Others 6.  This was very similar to their December poll.  

By this stage no-one with a clue should have found a Newspoll at 51-49 (ALP 31 L-NP 39 Greens 12 ON 7 Ind/other 11) in any way remarkable (other than it being the first ever released partway through an Australia Day long weekend).  However its release triggered a fairly large outbreak of poll denying on Twitter/X with some even spreading to Bluesky, which had previously been largely immune to such nonsense.  

My last-election prefs aggregate continues to give better numbers for Labor than the headline figures of most of the polls, with headline-figure based aggregates now clearly into the 51s for Coalition.  Some of the difference is respondent vs last-election preferences but a fair bit of it is also a run of polls where my 2PP estimates off the published primaries are better for Labor than the last-election/previous-elections preferences pollsters are putting out - even if the difference is in some cases only rounding.  It might be that the pollsters are sitting on something in the fine breakdowns that makes their 2PPs more accurate than mine but I've had reason to be unsure what methods some are really using, and distrustful of the calculations of some others.  If anything I would expect the low UAP vote between elections should push last-election 2PPs based on unpublished breakdowns in the opposite direction.  

What Would It Mean At An Election?

I believe my One Nation adjusted figure (based on preference flows from One Nation at the Queensland state election and the Fadden by-election) is probably a better predictor of overall preference flows than the 2022 election, so on that basis for "nowcast" purposes I'm using a 51.1-48.9 2PP to the Coalition, which is very similar to the current flood of 51-49s.  A lot of people want to know who wins at 51-49 to Coalition, but beyond that it would most likely be a hung parliament, the answer is it's anybody's guess.  

Following the Newspoll 51-49 there have been various pronouncements about what it all means, from claims that the Coalition is in an election-winning position through to points that by uniform swing that's only seven seats on the pendulum which isn't remotely enough for Dutton and co.  But both these positions are overconfident.   Firstly while there are only seven Labor-held seats below 3% on the pendulum (eight including Aston by its last-election result or my hybrid method) there are twelve between 3.1% and 6.1%.  That asymmetry means that if the 2PP lands around 51-49, by random variation alone Labor is expected to lose not seven seats but nine or ten, including Aston.  On paper any national swing against Labor gives Macnamara to the Greens (though there's not a lot of faith out there that that would happen) so Labor can easily be kicked down to 68 or so, with the Coalition similar.  If one also assumes that the outer-suburban uneven swing scenario in the Accent/Redbridge MRP is true, then 51-49 to Coalition can be something like 71 seats to 65 and just about enough for the keys to the Lodge.  On the other hand, those uneven swings might not actually happen, and the third round of teal attacks could inflict damage in Bradfield at least, so 51-49 might be a total mess.  And if it is a total mess, that favours Labor continuing, because the Coalition and supportive crossbenchers need to nail some blue to the mast and vote Albanese out on the floor if he declines to go quietly.

Leaderships

The Newspoll was a nasty one for Albanese with his net satisfaction dropping to a very ordinary net -20 (37-57).  It also saw his lead as Better PM shrink to a generally inadequate three points (44-41; the indicator on average favours incumbents by 14.5).  Historically, there have been four cases of a PM being re-elected after a worse netsat than this: Bob Hawke in 1990, Paul Keating in 1993, John Howard in 1998 and Howard again in 2001.  But of these only Keating was that unpopular on election day.  Howard's last figure worse than that in 1998, a net -27, was polled ten weeks prior to the election (this Newspoll was 11 weeks out if it's on April 12) and the others were much further back in the term.  

Resolve also had shocking stuff for Albanese with a net -22 for "performance in recent weeks" for Albanese (33-55), his equal worst from anyone of the term.  Resolve had a net +5 for Dutton (44-38 with rounding) (it had had him there before, and Essential has had a +6) and had a five point lead for Dutton as Preferred PM (39-34).  Previously only Resolve had had Dutton ahead on this all term, and only ever by one point.  (A YouGov in the Wikipedia tables with Dutton ahead by eight was Queensland-only).  On the other hand, Essential surprisingly had Albanese jumping up to a net zero approval rating, the only time he has been ahead of Dutton (net -1) in any poll since September, or in any poll that isn't Newspoll since June.

Can Labor gain support?

It remains the case that if the polls are right the Albanese Government's polling position is still very winnable; indeed since I've argued that 51-49 to Coalition is more or less a tossup, any improvement from here should put them in the box seat - again, that's if the current polls are right.  And sitting governments often do improve, as do underdogs.  Many past governments would have been happy with this polling position at this point.

Aside from the continuing drags of cost of living issues and voter pessimism, what makes me hesitate to compare the Albanese Government's situation to all the previous recoveries is that at no point in its term so far (beyond a very sparsely polled first seven months) has this government meaningfully gained support.  As the graph above shows, at least from very early 2023 the government's 2PP trajectory has been overwhelmingly flat or downhill with negligible ups.  (I think the ups and downs in late 2022-early 2023 are largely meaningless because of very sparse polling at the time.)  This is quite different to 1993 when Keating had already led in late 1992 before losing the lead, or the three Howard wins or Morrison's in 2019 in all of which the government had shown that it could peg back Labor leads.  

A similar story applies to Albanese himself.  As noted, Howard had worse ratings than this ten weeks out in 1998.  But Howard's ratings would go up and down a lot.  When he became on the nose it tended to happen sharply and also didn't last very long.  On the whole Albanese's ratings have just gone down.  Usually the more popular/less unpopular leader wins (1993, 1998 and 2004 are exceptions, though in 2004 John Howard did have substantial Better PM leads) so unless Labor can succeed in boosting Albanese or damaging Dutton they could have a problem. 

At the moment this is still a very open election.