Well last night’s referendum results turned out to be decisive. At the time of writing, No is sitting on 60.3% of the national vote, and has won all six states as well as the Northern Territory. The ACT voted Yes, and Yes is leading in just 33 out of 151 federal electorates.
The polls did a good job of picking the result, at least at a national level.
For this post, I’m going to focus on the variation in the vote across the country, and how it matches up against other electoral and demographic data points.
There was clearly a strong partisan element, but the Yes vote also did best in Inner Metropolitan electorates. I suspect this result is further confirmation of the realignment of inner city areas away from the Coalition.
The Voice referendum won’t go down as the worst-performing referendum, but it’s roughly in the bottom quarter. It is the twelfth out of 45 referendums to lose 6-0. At the time of writing, eight other referendums recorded a higher No vote, but it’s possible No could overtake two of those results.
There’s a strong geographic trend in these results. The AEC classifies electorates into four categories: Inner Metropolitan, Outer Metropolitan, Provincial and Rural.
A majority of seats in Inner Metro regions voted Yes, along with two provincial seats (Newcastle and Cunningham) and five outer metropolitan seats.
As a percentage, the Inner Metropolitan region voted narrowly Yes and the rest of the country voted No, with the margin increasing as you got further away from the cities.
This is obvious in my map, which show the strength of the vote in each seat as of Sunday morning.
There are some interesting conclusions I found when comparing the Voice results to other data points.
Firstly, I compared it to the two-party-preferred vote in 2022:
There is quite a substantial correlation, much more so than there was for the 1999 Republic referendum. It’s about in line with one of the less correlated referendums in 1988, the last time there was a referendum proposed by a Labor government and opposed by a Coalition opposition.
As of the time of writing, twenty out of 78 Labor seats voted Yes. This includes all three Labor seats in the ACT, ten seats in Melbourne, three seats in inner Sydney and the seats covering central Perth, Newcastle, Wollongong and the southern Tasmanian seat of Franklin. All seven urban teal seats voted Yes, along with all four Greens seats and Andrew Wilkie’s seat of Clark.
Only one Liberal seat is currently on the Yes side – the teal-adjacent seat of Bradfield in northern Sydney. Yes is narrowly behind in Deakin (49.2%) and Berowra (47.1%) but otherwise has not managed above 55% in any Liberal seat.
In contrast, seventeen urban Liberal seats voted Yes in 1999. A large part of the story is that most of those seats are no longer in Liberal hands, which is partly about the Liberal loss of the inner cities, but also because 2022 was generally a better election for Labor than 1998.
If you compare last night’s results to the Republic referendum in 1999, you get a much clearer relationship:
The Yes vote last night was about 5-6 points lower than the Republic vote, but the relative position of seats is very similar.
This pretty clearly shows an axis in Australian politics which has some relationship to the two-party-preferred vote but also diverges.
You’ll find similar correlations if you look at metrics like:
- Education levels
- Proportion of women in the workforce
- Young people and renters
- Support for the Greens (outside of a depressed Greens vote due to a local teal)
This Markus Mannheim chart shows the relationship between education and vote.
Here's a closer look at the associations. Caveats: these data are aggregates of large areas (federal electorates). The ecological fallacy may be wreaking havoc. pic.twitter.com/Pfy0gz5nzc
— Markus 🎲 (@MarkusMannheim) October 14, 2023
In a previous post I looked at how the 1999 Republic results lined up much more neatly with Labor’s 2PP in 2022 than Labor’s 2PP in 1998. The election voting trends now are more closely lined up with this axis than they were 25 years ago.
You can tell a story about how the Liberal Party lost so much of its support in the inner cities which starts with those seats voted Yes to the Republic in 1999, at a time when this vote cut across election voting trends. There was tremendous speculation about the Liberals losing support on the north shore of Sydney in the 2000s (remember the “doctors wives”?) that didn’t pan out, and then in 2019 and 2022 the Liberal Party lost seven seats to teals along with two seats to the Greens and at least one seat to Labor in the inner cities.
Last night’s result feels like a confirmation that those areas are out of step with the Coalition. That’s not to say the Liberals can’t win any of them back, but I don’t think that will help.
Labor has a challenge in holding together the two parts of its support base: the seats that voted strongly Yes and those that tended to vote No (although not by as much as Coalition seats voted no). But I don’t think Labor will need to be particularly worried about losing its marginal seats because of this referendum. I suspect it may be more of a problem for Liberal candidates trying to recover ground in the inner cities.
This is my only post for today. I’m going to be recording a podcast this afternoon and will hopefully have that up for tomorrow, and I’m planning some other analysis in the next few days.
In my seat of Macquarie, the Mountains/Hawkesbury divide is in stark focus yet again. Excluding the PPVCs and hospital polling, all polling places that had a 50% YES or greater vote were in the Mountains. 3 BM booths did not have a YES majority but were close – the lowest Mountain booth was 48.41%. The highest Hawkesbury booth was Bilpin with 45.66%. The lowest was Oakville with 17.64%. Katoomba had the highest YES vote at 75.89%. (numbers from this morning’s download).
Personally, I am not sure you can read too much politics into the result, other than a clear indication / correlation between wealth and these people being more progressive on social issues. Another issue, such as economic based issues, will possible draw a completely different result. The “others” are more likely to be disengaged, as they most likely had no / little skin in the game and therefore indifferent as to how they voted.
But I have to re-iterate a point I made a few days ago, the Yes campaign (political and organisational) ran a terrible campaign. As a supporter and putting on my critical thinking cap, I heard no compelling arguments to vote Yes. But maybe that (critical thinking) is not the thing in modern day politics and it is all about emotion and sloganeering.
Neil Flanagan: Critical thinking was alive and well and caused the No vote. For the critical thinker the biggest question went unanswered – it was where and what is the detail of the Yes argument? We were left guessing the answer which on the basis that BECAUSE it was MISSING despite the repeated requests for the answer, we were suspicious of what was hidden and rightly so given the way Albo flanked around the issue. The answer was necessary for a critical thinker who didn’t want to blindly trust Albo to formulate the detail after the referendum, which was his plan.
@ Neil Flannagan
I agree with you. I dont think you can read too much on it. I still think more people focus on economics than social issues especially during a cost of living crisis. Furthermore, we dont vote how lead us at a statewide level so an 90% no vote in some rural towns makes no difference when these are already safe areas rather it more important to win votes in the right places.
I think the strategy of Albo/Yes Campaign was based on the trauma of the 1999 Republic referendum. Many Republicans believe Howard had sabotaged the Republican case by providing too much detail on the referendum question rather should have simply asked if they wanted a Republic or even better for them just state an Australian Head of State. Therefore they believed an emotional campaign with the Vibe was the strategy. In hindsight, the John Farhnam ad may have sounded great in trendy inner city areas but just did not cut the mustard on Smith Street, Melton. Kos Samaras said reaching out to CALD community would have been critical especially in NSW/VIC many of these voters would not have been ideologically opposed but just need targeted information. In outer suburbs where people work two jobs to make ends meet they just did not have enough people to volunteer and explain the details. The ad with an indigenous child ask if he would live as Long as other Australians only came out towards the end of the campaign.
It’s possible to read too much int o arguments about the campaign – there is a history of mid term campaigns by labor Governments not being successful whatever the issue – most of the suggestions of what the Yes campaign should have done would not have made more than a marginal difference once the LNP decided it would oppose the proposal. Australians are reluctant to change the constitution – I suspect we now have a constitution that is frozen.
I’ve heard about this “sabotage” by Howard of the republic referendum often by supporters of the republic, with particular focus on putting the model in the question, but given referendums must be answered with either “Yes” or “No”, if the referendum question was just wanting to be a republic without the model and assuming it was successful, wouldn’t a 2nd referendum have to then put a model forward? Then you’d go back to the issue of some republic supporters not liking a particular model and willing to vote no as a result as was the case in 1999, and with a no result successful on whichever model put forward, then wouldn’t the status quo of the constitutional monarchy remain?
Labor should find a section of the constitution that is universally agreed needs updating and cut a deal with the L/NP to agree an amendment and put that to a mid term referendum in 2026 – Section 44 (Dual Citizens) would be a good candidate. It would smoke the Coalition out as to whether they would ever support a Labor proposed referendum proposal. Even if it is seems as beneficial the major parties (when it wouldn’t – as it stands it is a massive administrative burden for parties and the major parties are better able to deal with it than the minors and IND) – it would probably still get up if there was no major party opposition.
Regional Australia overwhelmingly voted No, including electorates with large indigenous populations e.g. Parkes, Lingiari, Kennedy, Leichhardt. However, within some electorates there are remote indigenous-majority communities such as Palm Island (Herbert), Cape York and Torres Strait Islands (Leichhardt) and remote polling stations and Jabiru in Lingiari. These polling stations were overwhelmingly Yes. This differs to larger No-voting towns like in the far west of NSW or regional QLD where indigenous populations make up a relatively large percentage but nowhere near half.
@Glenbrook Local, you made a good point about stark contrasts within electorates. There is a strong correlation between teal/Green votes and Yes-votes. Electorates with strong Greens-voting booths like Byron Bay in Richmond and the Blue Mountains in Macquarie were at least 60% Yes-voting at most booths.
In the seat of Lingiari, it seems the vote is very racially polarising with most Indigenous tribes having a heavy Yes Vote and non-Indigenous people having a Heavy No Vote. I think factors include that Lingiari would just be like Maranoa (if there weren’t a lot of indigenous) and the recent crimes reported in the NT, especially around Alice Springs might also been a factor in why non-Indigenous people voted No
Hi Ben,
I would like to raise a couple of points with you:
1 “Yes is narrowly behind in Deakin (49.2%) and Berowra (47.1%) but otherwise has not managed above 55% in any Liberal seat”: It should be above 45%, not 55%.
2 As for the chart “Voice referendum results compared to 2022 Labor 2PP”, you said there was “quite a substantial correlation” between Labor 2PP in 2022 and the Voice Yes vote, but the graph shows the correlation looks quite weak. Could you please plot the regression line so that we can see how strong or weak the correlation was? Besides, the difference between 2022 Labor 2PP and the Voice Yes vote for each electorate would be much more significant. Many urban safe Labor seats recorded Yes vote more than 20% below their 2022 Labor 2PP, and the safe Labor seats of Kingston and Spence in SA recorded Yes vote more than 30% below their 2022 Labor 2PP. Besides, you said there was “quite a substantial correlation” between Labor 2PP in 2022 and the Voice Yes vote, but the scatter plot shows the correlation looks quite weak. Could you please plot the regression line so that we can see how strong or weak the correlation was?
3 “But I don’t think Labor will need to be particularly worried about losing its marginal seats because of this referendum”, Labor should definitely worry about losing some marginal Labor seats that recorded the highest No vote among Labor seats, in particular Hunter, Paterson, Blair and Lyons. These seats will all be Coalition targets at the next federal election.
Hunter has a first term Labor MP who could benefit from a sophomore boost and has also never been won by the Coalition, therefore the latter three will be of more serious concern for Labor. The upcoming redistributions could also weaken Labor’s hold on Paterson. If Labor wants to retain a parliamentary majority after the next election, it has to defend these seats and other marginal Labor seats or win seats from the Coalition elsewhere.
@ WL
Yes many Republicans today says their was sabotage by Howard with the model put forward but that is retrospective justification and was not raised prior to the referendum. The model put to the 1999 referendum was not chosen by Howard rather it was the majority of the Republican delegates at the 1998 Constitutional Convention where Monarchists actually abstained when various models of the Republic were put to a vote. I agree if we had a non-binding plebacite on a Republic where a majority of voters said yes and then the voted no at a binding referendum to the model we would still be a Monarchy which is why i oppose any suggestions of that nature. The accusation by Republicans leading up to 1999 was different was not about the model rather the question on the ballot paper should not have been brief.
Joseph, I think it’s a relative question. You may say it’s a weak correlation but it’s much stronger than in the 1999 Republic referendum for example. You can find the data and do your own charts if you wish.
Ben: My final thoughts here are that this result & the above analyses further shuts out the prospect of a return to Liberal Government in my lifetime. The teal seats & possibly Bradfield in Sydney are gone for a generation. The passion (& now anger) in this referendum felt by yes voters in those areas was intense. Ditto in Melbourne & results in SA also reveal this trend. Where does Dutton win metropolitan or urban regional seats? Possibly by the next election the Nationals will be the Opposition unless the Qld LNP (or NLP) decide to vote as one party. Working class urban seats simply will never vote national.
I must admit, I am not so sure this leaves the Liberals (well the Coalition at least) out of power for a generation. If the Teal seats are gone, and I believe like you they are, then that leaves a very different Liberal party, and one which is more likely to find common cause with suburban voters. Just looking at the scatterplot above, there are 5 or 6 seats around the 50% 2PP and 40% Yes, and another 5 around 55% 2PP and 30% Yes. That means there are already 10 seats that on the same basis are already ‘in play’, and the likelihood that some others, some on what seem insurmountable margins, will also be in play.
I think the key thing to remember is the vote fell along socio economic fault lines, and the fact that the ALP put up a constitutional amendment only supported by those in a ‘higher’ (for want of a better word) class won’t go unnoticed.
The voice referendum won’t have as much long-term impact as the Brexit referendum did on subsequent elections and political landscape in the UK.
From memory, the Voice was not ranked highly on voters minds at about 2 to 3%. This meant that there is a lot of apathetic voters and likely voted No. The Voice is certainly a bigger issue in inner-city Green/teal voting suburbs, and chances are, those voted Yes.
In teal seats where the Yes vote was higher than the teal 2PP vote e.g. North Sydney, Wentworth, Kooyong, you can almost write off the Coalition in 2025, as issues like cost of living and the culture wars won’t cut through with voters as much as say, in the outer suburbs.
I wouldn’t be surprised if Julian Leeser faces a preselection challenge (depending on how active the hard right is over in Berowra). It holds a safe seat and rebeled against his leader. His electorate will likely expand south for the redistribution and he may have to battle either Alex Hawke or Paul Fletcher. If I heard Antony Green correctly last Saturday, Paul Fletcher will face Kylea Tink post-redistribution.
As for Bridget Archer, the other lower house Liberal who was a Yes campaigner, since she holds an ultra marginal seat, she is quite safe from a preselection challenge. Who would want to challenge for a marginal seat? Not unless there’s an imminent Coalition landslide victory.
Agree Votante
I dont think the Voice referendum will have as much of an impact Brexit had a much greater economic dimension that the Voice. Blue Collar & White Collar works perceived the effects/merits of Brexit differently but a lot of the arguments on both sides were economic rather than cultural. The Voice has been linked to better socio-economic outcomes for indigenous people but that does not impact the of non-indigenous people. The challenge for Labor now is to focus on bread and butter isssues so there is an economic narrative come May 2025 but still work on improving socio-economic outcomes for indigenous people. No one left, centre or right is argue for the status quo with respect to socio-economic outcomes for indigenous people. The only advise i would hope they take is not to focus on the republic for sometime.
@ Nimalan With the defeat of the Voice I suspect the Republic question is dead again for the foreseeable future. The Government needs to focus on the bread and butter issues as you say and I don’t think the polling numbers are as convincing as they need to be for it to be brought up anytime soon.
On top of that, I wasn’t alive during the 1999 referendum so I could be wrong, but despite Howard being PM at the time I get the feeling the current climate in the Coalition as a whole is more negative towards a Republic than it was back then.
It’s an issue best left for late this decade or in the next one.
I suspect that it is mainly the highly educated white-collar votes that voted Yes but the ultra-wealthy (regardless of education) voted no indicating that is more professional class rather than the business class that votes more progressively. This can be evident with Toorak being the richest part of Higgins but was the only booth that voted no.
Anyone have any idea how much of the vote is still to be counted? Seems to be around 80% counted? If this is the final turnout it is woeful. Much lower than elections and previous referendums. I suspect all the effort the AEC went to to increase enrolment went to waste as they didn’t turnout.
I think the republic referendum should be shelved for at least a term, possibly two. The prospect of referendums is dead for now as it could be politically risky to be seen as taking the eye off the ball. Add to that, many of the voice debate arguments could be used for the republic debate. There are some recurring themes. For one, the model of a republic could cause infighting and side-switching, similar to the voice debate did. It is also prone to misinterpreation.
A good time to get the republic done was when Malcolm Turnbull was PM so at least they could’ve gotten bipartisan support. The problem was that he would’ve upset the hard right and he would have to sacrifice his career. It was also when there was the same-sex marriage plebicite was going on and that would’ve given voters ‘electoral fatigue’.
I do think there is more impetus for a Yes victory than in 1999. There’s a growing non-Anglo-Celtic population. There are immigrants from countries that don’t view British colonialism that kindly e.g. India, Pakistan. India has abandoned the old colonial names of many cities and states and renamed them. On the non-demographic side, King Charles isn’t as popular as Queen Elizabeth. Many republicans willingly waited till after she had died to start the republic campaign. Since Charles ascended at 73, he carried a lot of baggage into his new job from royal family scandals to the way he treated Princess Diana to the amount of taxpayers dollars he gets.
@Ben Raue I noticed you make these Leaflet/OpenStreetMap maps. Could you please make some .map files for each electorate? I’ve been trying to make some but they don’t work.
What’s a .map file?
@ Laine/Votante
Agree that the Republic should not be revisited until 2030 at the earliest. The Chance of Bipartisanship are less with the hard right having more sway. Even if there is a moderate Liberal as leader in 2030 it may not a battle they are willing to fight their own party on especially if the Hard right are willing to blow up the show and prevent moving towards a pragmatic climate policy as well. Also i feel the strongest Republican vote will be in the more bohemian Green areas rather than Teal so for the Libs it better to focus on climate, getting more professional women and economic differentiation as there is really no votes to pick up on this issue. Mind you i dont think Labor can win many votes from the Greens or Teals by championing a Republic.
High Street
Great idea for a referendum.
S44 a persons eligibility to become a Federal MP, should be based on Australian citizenship, irrespective of other nations citizenship laws/customs/history.
Actually, I think a good idea for a referendum is to have one so that referenda can only be held at the same time as elections for either house?
Given the difference between NSW and Tasmania – it is extraordinary that the yes/no vote in both states is within .02% of each other.
TAS: y – 41.06%, n – 58.94%
NSW: y – 41.04%, n – 58.96%
Without checking this has to be the closest vote between states ever. And certainly for NSW:TAS as they are at the opposite end of the Federation in relation to size, income, demographics (age, religion, and heritage) and just about any other metric that you can think of.
Very surprising (almost identical) vote.
Best Pollster
Few people on this blog comment from a culturally conservative view, so I will offer one here.
There is virtually no appetite for a republic within the Liberal Party now or at any stage in the future. Republicanism of any kind is far, far less common within the Party than at any time I can recall, and I have been a member since 1993.
To me, the assumption that migrant communities will move the dial on the issue doesn’t seem to hold up when looking at attitudes among Italo-Australian and Greek-Australian attitudes towards a republic over recent decades. The lesson of the Voice referendum is that people continue to follow the iron laws of referendum physics, regardless of their surname. So it would likely be with any future republican re-run.
Contemporary republicanism is dead. Look forward to King George VII one day.
@ Hughie
I think you bring some excellent points.
1. I agree the Libs are far less interested in the Republic than in the 90s or early 2000s
2. I actually think the Republic may actually underperfom these days than in 1999 in the Teal demographic seats (including Bradfield, Boothby, Sturt, Higgins). In nearly all those seats the Voice actually underperfomed the Republic. The one exemption was the Northern Beaches where Warringah and Mackellar had a better result for the Voice than the Republic in 1999. So i dont actually think constitutional change is a major driver for the small l liberal vote.
3. In fact it was the inner city bohemian hard left areas where the Voice outperformed the Republic (Melbourne, Gryandler, Cooper, Wills, Sydney) and parts of Newcastle, Wollongong, Geelong etc. These areas are not really a focus for the Libs anyway.
3. I do agree that many ethnic communities especially the more established ones such as (Italians, Greeks, Croats, Jews, Lebanese Christians, Other European Communities) are better integrated and now well into the third generation so maybe less likely than the older generations to vote for a Republic.
4. I also think the Republic is strongest among Late Baby Boomers/Gen X, who remember the Whitlam dismissal etc and never got much traction among Millennial/Gen Z has not been discussed for years at the grassroots level
5. I feel its is actually Green voters/ Vic Socialists etc that are most passionate about the Republic and not really Labor primary voters.
Comments are closed.