With a hung parliament looking very likely at the upcoming federal election, there has been quite a lot of looking back at the last hung parliament in 2010.
Amongst other elements, there has been some parallels drawn between the experience of the crossbenchers in 2010 and how the current crossbench could jump if they find themselves in a position to decide who forms government this year.
Narrowing down further, it is worth drawing a comparison between Tony Windsor and Rob Oakeshott, two former Nationals members representing traditional Nationals seats who supported the Labor government, and those crossbenchers who represent traditional Liberal seats today – the “teals” and some similar members representing regional seats.
Windsor and Oakeshott ended up retiring at the 2013 election. While their electorates overlapped with independent seats at a state level, all of those members were gone by 2013, and independents haven’t been competitive in those areas since then. While we don’t know how these two would have performed at the 2013 election, it is likely that they would have been defeated. Other cross benchers did not suffer the same fate. Left-leaning crossbenchers Adam Bandt and Andrew Wilkie represented traditional Labor seats and while there was a swing against the Greens in 2013, Bandt and Wilkie both strengthened their position that year and still hold their seats. Likewise Bob Katter’s decision to support a Coalition government in 2010 has stood the test of time.
While there is now a larger cohort of teal MPs representing traditional conservative seats who may find themselves deciding who forms government after the election, I don’t think their position is comparable to Oakeshott and Windsor. In this post I’m going to examine some electoral statistics to show how their electorates, and their voter bases, differ from those of the 2010 independents.
So first it’s important to clarify what statistics we are using.
When votes are counted for the House of Representatives, they are first collected into primary vote piles. Then the votes are distributed, with one candidate knocked out in each round.
The final preference count is called the ‘two-candidate-preferred’ count, and in all of the seats I’ll be discussing today that final count is between a Liberal and an independent. But there is also a two-party-preferred count, which takes all the votes and distributes them between Liberal and Labor.
Finally I will also reference the three-candidate-preferred count. This is the second-last round in the preference count, and in all of these seats was Liberal vs Labor vs Independent. You can thus compare it to both the 2CP and 2PP to identify how many voters ranked Labor, independent or Liberal in which order.
I should also note that I will be mostly focused on the six newly-elected “teal” seats from 2022, plus the seats of Warringah, Mayo and Indi. All nine have some commonalities. They were Liberal seats before being won by an independent. The Liberal Party remain the runner-up in the seat. The sitting members are all broadly people who sit to the left of the Liberal Party with socially progressive positions. I will sometimes use redistributed data, in which case I will include Bradfield in place of North Sydney.
I should also note that I am treating Rebekha Sharkie as an independent. While she remains a member of the Centre Alliance, that party has ceased to exist outside of her candidacies. Later on I will also examine Bob Katter, and I will likewise treat him as an independent even though he has run for his own party since 2013.
The first thing to note about these seats is that they have been shifting to the left in recent years. This first chart shows my estimate of the two-party-preferred vote (ie Labor vs Liberal) for these nine seats at every election since 2004. These figures have been adjusted for the current 2025 boundaries, and have been subtracted from the national vote. A ‘0’ would indicate that the seat’s 2PP was exactly the same as the national 2PP. Any lower number indicates the Liberal Party did better in that seat.
These seats used to lean quite far to the right. From 2007 to 2016, these seats voted about 11-12% to the right compared to the country, on average. This was less dramatic in 2004, but that was an election noted at the time for the fact that Labor gained swings in inner city areas while losing ground elsewhere. This is when the term “doctors wives” was coined.
Bradfield and Curtin were almost 20% more right-leaning than the country in 2016.
But in 2022, none of these seats leaned more than 10% to the right. Indeed Mayo gave Labor a 2PP majority in 2022, voting just 0.5% to the right of the country.
It’s worth noting that even if there is a swing to the right in 2025, polls are suggesting that these seats will not swing as much to the right as other seats. The first YouGov MRP this year showed these nine seats swinging to the Coalition by 1% on average on the 2PP, while the rest of the country was swinging 3.4%. The second MRP shows those numbers down to 0.8% and 2.1%.
So seats that were once heartland Coalition seats are rapidly coming closer to the national median.
So if you were an independent MP trying to decide if you were going to support a Labor or Coalition government, you’d want to know who is in your voting base and how they might lean.
To start with, this graph shows a breakdown of those voters who ended up in the 2CP for one of these independent MPs, showing their original primary vote.
None of these independents is close to winning a majority of the vote on their own. The vast majority of preferences they receive come from either Labor or the Greens – both because Labor and the Greens make up the lions share of preferences available, and because they tend to preference these independents at high rates.
Another way to look at that data is to look at how votes from a particular party flowed as preferences across these nine seats. Labor, Greens, One Nation and the UAP ran in all nine seats.
Labor and the Greens made up over 70% of available preferences, and both parties gave independents over 80% of their preferences.
But there is a different statistic that makes the preferences of these voters even clearer.
By comparing the 2CP, 2PP and 3CP, you can calculate exactly which proportion of independent 2CP voters also gave Labor their 2PP.
For this chart, I’ve also included the three Queensland Greens seats, where they were against the LNP in the 2CP. I’ve also included Kennedy in 2022, and Lyne, Kennedy and New England from 2010. The comparison is stark.
For the nine seats I’ve been discussing up to this point, the proportion of independent voters who preferred Labor to Liberal is overwhelming. The proportion ranges from 71% in Indi to 82.7% in Kooyong. This is not quite so high as it is in the Greens seats, where well over 90% of Greens 2CP votes prefer Labor over LNP, but it’s not that far off.
In contrast, Kennedy, Lyne and New England look very difference, with roughly half of voters for those independents preferring Labor to the Coalition. Indeed in Tony Windsor’s seat in 2010, barely 43% of his voters preferred Labor.
For one final graph, this one simply compares the 2CP and 2PP in each seat.
The Labor 2PP in the current independent seats is much higher than for the independents in 2010. It only exceeds 50% in Mayo, but it comes close in Warringah and North Sydney.
None of this is to say that these independents solely rely on progressive voters to get elected. Generally they need a large base of progressive voters to get close to victory, and then need some conservative voters to push them into majority. But in the current batch of independent seats, that base of progressive voters is much bigger and that extra push of conservative voters is much smaller than in Lyne or New England in 2010. The direction of travel makes it quite plausible that more of these seats could see Labor winning a majority of the 2PP in the near future. That was never close to a possibility in New England or Lyne in 2010.
When you see those different figures, it’s clear that the experience of the rural independents in 2010 doesn’t have much to tell us about what is happening now.
I’m not saying these numbers mean independent MPs should definitely support a Labor government. A lot will depend on the exact numbers in the parliament, what deals are offered and the nature of the negotiations. But I don’t think people should assume that independent MPs representing traditionally conservative seats would by default favour a Coalition government. Those seats are changing, the progressive bloc of voters in those seats has grown significantly in recent years, and those voters make up a key part of the independent voter base.
We will be able to update these statistics following the election, but not all of the data will be available quickly enough to be useful in the case of a hung parliament. So I hope this post is useful in assessing how these seats sit in the political context.
The missing piece of this analyst is the Senate votes which indicate how teal voters would vote when there was no teal on the ballot. I did a limited analysis of this in 2019 in Warringah and if my memory serves me correctly the coalition vote was about 5 to 15% higher for the Senate than for Abbott.
The impact in 2022 was about 0 to 10% in Warringah.
I did look at the Senate, I just didn’t think it added much.
Around 10% of people in each of the teal electorates votes 1 Teal 2 Liberal and 3 Labor: https://youtu.be/mEFIar4Kazk?si=mG9J9jIPdwxNd6ZI&t=327
Well to be more exact, about that number ranked those three options in that order. They may have voted 1 for someone else.
It varied from 9.2% in Kooyong to 14.7% in Warringah, and about 16% in Mayo and Indi. In contrast it was 30% in Kennedy in 2022, and about 40% in New England in 2007 and 2010.
There’s also a number of voters who put Labor, then Liberal, then independent. That varies from 2.1% in Kooyong to 5.5% in Mayo.
I think this comparison doesn’t quite work because you are comparing two similar candidates. I would think the Teals are more to the centre than Labor. Think of it as if there were a 3 cornered contest between Coalition parties, if Nationals were 1, Liberals 2 and Labor 3 it wouldn’t mean that Labor was anywhere near ever winning the seat. If a centralist were not running, the voters would go back traditional voting patterns
I think 3rd party/independent candidates can be divided into two broad categories.
The first is the popular local politician who is running against the big two, it could be Ted Mack, Dai Le, Oakshot, Windsor or someone else. The second is the “third way” (whether a party or independent). They might be DLP, Australia Party, Australian Democrats, Greens or Teals.
When Australia became a nation, the two major parties were the “Free Traders” and “Protectionists” perhaps reflecting the two sides of the economic debate. Then, for about 100 years it was Labor vs various conservative parties reflecting the bosses vs workers. In recent times, the economy has moved past the class wars to now be predominantly development vs conservation. The Greens had the opportunity to move to the centre and become the party of the left but they instead moved to the left of Labor embracing issues that are more to do with student politics than middle Australia.
The Teals will remain a force until they need to stop sitting on the fence. For the Australian Democrats it was a decision of whether to back the GST or not. For Oakshot/Windsor it was whether to back Labor or the Coalition. Once the “third way” candidates side with one or the other, they lose their support from their supporters who believed they would come down on the other side.
I don’t understand your first paragraph, David. What comparison? The teal comparison to 2010? I don’t understand the reference to a Liberal-National contest.
The point is that voters aren’t putting Labor third, most of the teal 2CP voters put Labor ahead of Liberal.
I do think it will be challenging for the teals if they have to “get off the fence”, but I think the assumption that it would be supporting Labor that would be the biggest problem is likely wrong.
The whole point of the data is that their supporters aren’t evenly split.
Ben
If you are voting for a sitting MP – of whatever persuasion – and 95% of the time will end up being in the top two – do you really pay much attention to how you number the rest of your ballot? If I am voting for a minor party yes I do care about my preferences – for a major no I don’t. And I am probably not alone. In 2022 – I think I voted Labor 1 then Liberal 2 and probably donkeyed the rest. My preferences beyond 1 were forever meaningless. The high notional ALP vote in Mayo and Warringah was because the public knew who the top 2 would be and didn’t care after that. And both Sharkhie and Stegall came out with high margins. It needs those two occurrences – incumbent non major and high margin.
In 2022, most teal voters were ex-Labor or ex-Green voters voting tactically, rather than disgruntled lifelong Liberal voters. It makes sense that most teal voters put Labor ahead of the Liberals as per the third chart.
https://www.abc.net.au/news/2022-12-05/election-results-teal-independents-anu-study/101732654
There is a possibility that Liberal-turned-teal voters (some, not all) numbered Labor ahead of the Liberals, not because they liked Labor or Albo, but because they felt disappointed by the Liberals or wanted to punish them.
Interesting figures Ben. I would like to ask where did you get the figures of the proportion of crossbench voters who preferenced Labor over the Coalition from. AEC does not publicly release two-party-preferred preference flow by candidate for non-classic seats. And why are your figures so dramatically different from Kevin Bonham’s? For example, in Curtin, according to you, 77.8% of Chaney voters preferenced Labor over Liberal, but according to Kevin Bonham only 67.8% of Chaney voters did so. In Mayo, according to you, 74.1% of Sharkie voters preferenced Labor over Liberal, but according to Kevin Bonham only 60.2% of Sharkie voters did so. https://kevinbonham.blogspot.com/2022/07/2022-house-of-reps-figures-finalised.html?m=1
You’re right, they don’t publish 2PP preference flows for primary vote. The only one of these seats where we can calculate that stat is Mayo, because Sharkie was the only CA candidate.
I have calculated what share of the independent 2CP that flowed to Labor or Coalition. It looks like Kevin’s calculations are the share of the independent 3CP. So my number will already be higher, since he’s calculating the number before Labor’s preferences are added.
As far as how I calculated it, it’s not that complicated, it’s a combination of the 3CP, the 2PP and the 2CP. The difference between the IND 2CP and 3CP, and the LIB 2CP and 3CP, tells you the number of voters who voted ALP-IND-LIB, or ALP-LIB-IND. The difference between the ALP 2PP and 3CP, and the LIB 2PP and 3CP, tells you the number of voters who voted IND-ALP-LIB or IND-LIB-ALP. The LIB 3CP can’t be split between those who preferenced Labor next or IND next, since there is no IND vs ALP 2CP.
If you add up the ALP-IND-LIB, IND-ALP-LIB and IND-LIB-ALP, that gives you the IND 2CP. The first two categories are those that are in the ALP 2PP.
I think Labor winning a majority of the notional TPP in a teal seat doesn’t mean the seat would actually vote Labor over Liberal. Warringah was close on notional TPP last election with the Liberals having the narrow edge (note the poor candidate choice) but on the state level Warringah is entirely within Liberal seats held by moderate MPs, and the demographics of the area aren’t the demographics of a Labor area at all.
Yes my calculations are the 2PP split of the independent’s 3CP (and also of the independent’s primary vote where I can do that) – I look at this because I want to get a handle on the question of who is supporting the independent? Are they left or right leaning voters?
One thing missing from your analysis Ben is Simon Holmes a’Court.
If there is a hung parliament, will it reveal the TEALs are truly independent, or do what Holmes a’Court says.After all, he has financed them all-so they are under some obligation to him.
He will not support the Liberals.
Even if I am wrong about this, I think it would be a mistake for Dutton to go into minority government.Although any odium will be shared, the minority government will bear the majority of the blame
Nether Portal
As a former resident of the Beaches (Warringah and Mackellar), I would say if the leftward trend noted by Ben continued it would be more likely to result in a Greens candidate being elected at some point (in the absence of a teal candidate), similar to the result in Ryan, rather than Labor.
Your characterisation of the state level dynamics isn’t quite right though – the state seat of Wakehurst which overlaps the northern part of Warringah is held by independent Michael Regan with teal characteristics, although he long predates them as he was previously the mayor of Warringah and then the amalgamated Northern Beaches Councils from 2008 to 2023. A teal won Pittwater (overlaps Mackellar) at a by-election last year, and Manly (entirely in Warringah) has a long history of independents. Although the Liberals’ nomination fiasco at the local elections last year means they’re currently underrepresented, there are 4 Greens and 7 independents affiliated with Regan on a council of 15
@sabena Dutton wouldn’t be able to govern in a minority and cross-bench that needs negotiation. He can only do his thing by having a majority and shutting down any debate
I think its fascinating that Ben puts so much effort into producing this data and the lovely graphs – that OVERWHELMINGLY displays the preference of Teal voters for Labor over Liberal (>75.8% in 6 out of 8 seats) -and so many posters simply say “Oh, you can’t rely on that.. the voters didn’t really mean it”.
And as for you Nether Portal, please enlighten us as to what the “demographics of a Labor area” actually are nowadays? And don’t say that Bennelong & Reid and similar seats don’t count – they are held by Labor so by definition they are ” a Labor area”.
@Ben Raue Thank you for your explanation. I initially though your third chart is showing the 2PP splits of crossbenchers’ primary votes (how voters who gave their primary votes to a crossbencher preferenced one major party over another) while it is in fact showing the 2PP splits of crossbenchers 2CP vote (how voters who preferenced a crossbencher over the Coalition preferenced one major party over another). The figures in the red area in your third chart are essentially (ALP-IND-LIB + IND-ALP-LIB)/(ALP-IND-LIB + IND-ALP-LIB + IND-LIB-ALP).
@bazza-you are simply putting my comment in a different way.Why would he bother-there is no upside for him.
I’d say that the vast majority of the primary vote for the Teals were from the Labor and Greens hence the crash in their primary vote in teal seats. Those went on to preference Labor ahead of the Libs which is why the majority of the Teal preferences would go to Labor in the TPP vs the Libs. However, it was the minority of the Teal vote which went from Lib to Teal that handed the Teals their victories hence the Teal seats still have the Libs winning the TPP against Labor.
It is simplistic but in most of the Teal seats you can see where the build up is. In Kooyong, the decline in Liberal, Labor and Greens votes from 2019 plus Oliver Yates 2019 vote is almost (0.2% off) a complete match for Monique Ryan 2022 at primary. Goldstein is similar. In Kooyong only 15% came from the right, Goldstein much higher. Other Teal vote you can see the buildup. The Teals have largely eaten Liberal seats but they have actually eaten a lot more Labor and Greens primary votes. That has impacts on funding, senate votes and local branch morale.
@NP, didn’t the Teals ironically made Labor only had a minor swing within a Libs vs Lab TPP (in comparison to nearby Lib vs Lab Seats) given some Teals didn’t put preferences on their How-to-Vote cards alongside many Teal Voters perceiving the no longer needing to preference Labor ahead of Libs.
Dan M, you’re correct. But I think it would be a mistake to assume only the views of that small number of ex-Lib voters should be decisive in working out what position to take. I’d say that most of the volunteers for these teal MPs would lean to the left, and there is even a danger they could slip out of the 2CP if enough of the left-wing base chose to favour Labor or the Greens.
This shows the dilemma that ‘teals’ and other ‘third way’ candidates are facing. While the majority of their voters may prefer Labor over Liberal, a majority of the voters in their seat prefer Liberal over Labor. Given they are in a 2PP vs Libs, logic dictates they should back a Coalition government, since the proportion of their voters who prefer Labor will still vote for them over Libs at the next election. In contrast, if they back Labor, the proportion of their voters who prefer Libs over Labor may switch back to the Libs, costing them their seat. However, as Ben points out, such logic is secondary to the election results, nature of negotiations, personal ideology and soforth. I would prophesise that currently the ‘teal’ policies are closer to Labor’s policies than the Libs’. If we do get a hung parliament, it will be quite the conundrum for them to resolve.
It might be possible that many teal volunteers and supporters are left-leaning and got behind teal candidates as they had a greater fighting chance compared to a generic Labor or Greens candidate.
Most teal seats in capital cites are wealthy, progressive and socially liberal as indicated by the same sex marriage survey and voice referendum results. Before 2022, there was a growing Green or tree Tory or “Liberals for Forests” vote or “Doctors’ Wives” vote in some areas e.g. Avalon Beach to Palm Beach in Sydney, Hawthorn in Melbourne. An ongoing political realignment of uni-educated professionals voting left may be part of it. What limited the growth in the left wing party vote was their economically conservative leanings which aligned more with the Liberal Party of Howard.
@Redistributed – very much agree, although it is lazy of commentators and analysts to simply lump all the Teal seats into one box. There are differences and they are not all safe Liberal seats – Wentworth certainly isn’t and hasn’t been for about 20 years. The Labor primary vote in North Sydney and Bradfield in 2022 didn’t drop much (about 3.5%) on the 2019 result.
The teals and most independents will not be swayed by the possibility of losing their seats. It is Possible they are joined by another 4 or 5 cross benchers
Mick,
Pull the other one.Every MP of whatever party is swayed by potentially losing their seat.
@Ben Raue – it is possible to ask you to do a little bit more work and work out where the 2022 version of Bradfield would sit on those last two graphs?
Thanks
For the last graph, the numbers can be found in the results table of my Bradfield guide. The numbers are 43.8% and 47.5%. The Labor 2PP is higher than Mackellar but obviously the independent 2CP is lower. I can’t do the other one without calculating 3CP which is a lot of work.
Mick, most independents are first-term MPs and also happen to have marginal seats on paper. There is little room for complacency and it shouldn’t be assumed that they are safe. Then there are the three coalition-turned-independent MPs (Gee, Broadbent, Goodenough) who are all running for reelection and there’s a possibility of losing.
Oh sorry that’s the redistributed version of Bradfield.
Pre-redistributed, the numbers are 43.5% and 45.8%. Which is obviously very close, but on the other hand that was an election-losing result, and I suspect the part of the teal base who preferences the Liberals are in that segment that wasn’t flipped in 2022.
For the 2022 version of Bradfield, 82.8% of Boele’s 2CP voters preferenced Labor over LNP, which is higher than any of the successful teal seats.
But again, I think you would expect that number to drop a bit if Boele was to win.
Ben, thank you for your hard work on this, it is really interesting.
One question- do you think it possible that the difference between the Labor 2PP and the Independent 2CP (as outlined in your graph above) could act as a proxy for the personal vote of each independent candidate? Or is that something only possible to see after the election, when we can compare two results across time?
I think it does have some relationship.
Thanks Ben,
are you saying that your view is that for Boele to win she needs to reduce the Liberal primary – by definition – and that this portion who has not (yet) flipped would be less likely to preference Labor if they did?
Even given the above, it is interesting that Bradfield, Kooyong and North Sydney were all basically the top of the list at 82-83%
Yes I think so.
At the very least I’m saying that you can’t compare a winning teal effort to a losing teal effort.