10 November – The AEC has carried out its final adjustments to the roll to take account of deaths, re-instated voters and several other causes of adjustment. This reduced national roll by around 5,000 voters and lifted the final turnout figure from 89.92% to 89.95%.
In this post I’m publishing several charts dissecting the referendum result by Vote Type and by electoral division. The post includes a table of Referendum Yes percentages and comparison columns for Labor two-party preferred percentage from the 2022 election, plus the gap between these two figures.
The electorate table shows how much lower the Yes% vote was in many traditional Labor seats. The seats where the Yes% was higher are clustered in seats won by Greens and ‘teal’ Independents at the 2022 election, and also several Liberal seats gained by Labor.
The ‘gap’ column shows a similar pattern to the 1999 Republic referendum. Both the 1999 and 2023 referendums saw Yes support distributed very differently from two-party preferred patterns at the preceding Federal election. The 1999 referendum pattern was also very different to the 2001 Federal election, which suggests the 2023 referendum is unlikely to be a guide to voting patterns at the 2025 Federal election.
That’s with the possible exception of the result in seats lost by the Liberal Party in 2022. Of the 17 Liberal seats that voted for the Republic in 1999, only five were won by the Liberal Party in 2022. The other twelve seats are now held by Labor, the Greens and ‘teal’ Independents. Eight of these seats voted for The Voice in 2023.
Result by Vote Type
The chart below dissects Referendum Result by vote type. The AEC Commissioner stated after a week of counting that the final turnout would be in the range 87-89%. The final figure pending roll reconciliation is 89.4%, down only slightly from 89.8% at last year’s Federal election.
The chart separates Ordinary Votes into Polling Day and Pre-Poll Ordinary votes. Postal votes have been counted and the final return rate for Postal Votes is 86.4%. All figures used here are as at 30 October after all votes should have been returned and counted. The final and official AEC figures may differ slightly once final checks and reconciliations are complete.
For someone like me who covers elections for a living, the most remarkable feature of the above graph is the eight percentage point gap between polling day votes and pre-poll ordinaries, and the even greater 11 point gap with postal votes. Given the greater volume of pre-poll and postal votes at recent elections, the emergence of such gaps greatly complicates the calling of election results. There is a traditional conservative lean in postal and pre-poll votes, but the lean varies inconsistently from election to election. A lesson from the Referendum result, and from the NSW election in March, is that calling the result of close elections is going to take longer than in the past. Especially close elections where Labor is leading narrowly.
The above chart also separately identifies four smaller categories of votes, Special Hospital Mobiles, Remote Mobiles, Other Mobiles and Provisional votes. Remote Mobiles includes mainly votes taken at remote indigenous communities, but also includes some remote mining and agricultural communities.
Referendum Results by Division versus 2022 Federal Election Results
The chart below shows referendum results by division in comparison to 2022 Labor two-party preferred results. Each line shows electoral division, state, holding party, Yes %, 2022 Labor 2-Party Preferred % and the gap between the two. (Gap = Yes% - ALP 2PP%)
(The data could have been shown with Coalition 2PP% and No%, but it is easier and more interesting to explain the smaller number of electorates that voted for change than the much larger number that voted for the status quo. The Coalition % and Nos % in the table below are simple 100 minus the percentage shown.)
I have identified several of the independent held divisions with the label 'TEAL' to aid clarity though no such party exists. As is clear from the chart, the so-called 'Teal' seats behaved very differently from other seats with an underlying Liberal two-party preferred majority. '*' indicates seats that changed party at the 2022 election.
You can sort the table by the columns across the top.
I find it remarkable that no one seems to be looking at the foreign-born demographic when considering the referendum results. I have two main circles of friends, both mainly university students or recent graduates. The Australian-born ones were, unsurprisingly, almost unanimously passionate ‘yes’ supporters. But every single one of the foreign-born ones, and their whole families, were strong ‘no’s. And considering the Australian electorate’s percentage of foreign-born voters is one of the highest in the world, that has to have been a major factor.
COMMENT: It might explain some variation in the cities, but country electorates had the lowest Yes vote and they have the lowest proportion of migrants. Strongly migrant electorates voted against same-sex marriage but the pattern at this referendum is different. Higher status occupation, higher education levels and higher income levels levels explain most of the variation on Yes and No vote, as they did in the Republic referendum. I think migration levels is a lesser factor than those.
I’m a migrant here, and I raised this issue on both the voice website and with SBS researchers. I flagged the lack of explanations on the nuances of the situation for those not familiar with it, those who aren’t native english speakers, migrants who come from countries that have been invaded for centuries, and also the lack of explanation on how it was going to work for migrants who don’t know how it works now. It’s also worth keeping in mind becoming a citizen here is a mission, and often takes years – I’m not sure how magnanimous individuals were feeling when they have received no government support for years on temporary visas after jumping through all the immigration hoops.
You’re right it would be really interesting to read more analyses on this.
Thinking about the polling day to prepoll ordinaries gap, is there variation between divisions, and if so, are there demographic factors that explain that variation? I might try to answer that myself since it’s an interesting phenomenon.
Only one Liberal electorate voted yes, Bradfield, and that seat is wealthy. It went from a safe to marginal seat in the last election, so is almost Teal. No National seat voted yes. This result must in part be due to the Coalition opposing The Voice. If bipartisanship was achieved, would that have made the difference? I think so.
Also, I know ostensibly good people with modest incomes and who do not have high levels of education. Some reasons they voted No we’re: The current circumstances of Aborigines is not my fault, they get free cars, are not punished for stealing carts of groceries, already receive too much Government funding and waste the money they do get.
Many of us are strongly individualistic (especially Liberal voters) and blind to the disaster of colonisation on many Indigenous lives today. We are ahistorical in our thinking. The truth is out there but a lot of us can’t or won’t hear the truth. The Voice stats are complicated to interpret. But a big part of the story it tells is that many Australians do not like Indigenous Australians, and lack a generosity of spirit about their suffering.
HI Antony, I understood that as the pre-polls and postals were counted, the NO% was likely to increase – and some commentators were quoted as predicting that the final result would be 62 to 38. But the reverse seems to be happening, and the NO %, after increasing for some time, thereafter declined, and now looks like it will slip below 60%. Any idea what’s going on here?
COMMENT: The No vote rose to 60.8% with the addition of postal votes on the Sunday and Monday after the election. Nearly everything else counted since then has been Polling Day Absent and Pre-Poll Absent, both of which traditionally lean towards Labor and the Greens. Both categories recorded a No vote, but much lower than for polling day, pre-poll and absent votes. As the Absents have been included in the count, along with later arrive postals, the No % has shifted down to 60.1%. I don’t expect any further shift.
Thank you. That’s a very useful table. One thing though, “Teals” are not a party or a thing, notwithstanding Liberal Party hysteria and NewsCorp lazy journalism. It is correct to classify them as Independents
COMMENT: If you read the paragraph above the table I stated that clearly. My reason for doing the separation was because seats held by so-called ‘teals’ had significantly different results than for seats held by other independents. Voting for ‘teal’ independents and voting Yes are clearly correlated results.
You’re right, but they are from a very important, high-earning type of seat from the perspective of the referendum, so it is easier and more informative to classify them differently here..