Australia’s Constitution remains frozen, with the loss of a referendum on 14 October 2023. Only eight out of forty-five national referendums to amend the Constitution have succeeded, with no successful change since 1977. As the periods between holding referendums get longer, the people become unfamiliar with the mechanisms of constitutional change and even more wary about approving it. This feeds a spiral of loss and increasing reluctance on the part of governments to spend their political capital on futile constitutional reform attempts.
The 2023 referendum would have recognised Indigenous Australians in the national Constitution and provided a means, described as a ‘Voice’, for them to make representations to Parliament and the Executive Government on matters relating to Aboriginal and Torres Strait Islander peoples.
Harry Hobbs has previously explained the mechanism for constitutional change and the background to the Voice referendum. In this post, therefore, I will seek to give an overview of the reasons for the failure of the referendum, including the ‘No’ arguments, factors that contributed to the ‘No’ vote, and the demographics of the voting outcome.
The arguments for ‘No’
Opponents of the Voice referendum came from the political left as well as the right. The ‘Progressive No’ campaign from the left, which attracted some support from Indigenous groups, regarded recognition in the Constitution as inconsistent with Indigenous sovereignty. It contended that Indigenous Australians should not engage in a ‘colonial’ constitutional process. It also portrayed the proposed Voice as a powerless and ineffective body, and instead campaigned for a treaty and a process of truth-telling, which it regarded as more likely to occur if the Voice referendum failed.
The Conservative No campaign’s most effective argument was that the amendment would be divisive and insert race in the Constitution. It argued that it was racist to include in the Constitution a provision that gives a Voice to a group of people based upon their racial ancestry. It claimed that the amendment would divide us by race by giving ‘special rights’ to Indigenous peoples that are not held by others, destroying the equality of citizenship established by the Constitution.
The proposed amendment was described as creating two different grades of citizenship, based on race. In effect, it turned on its head the argument that opposition to the Voice was based on racism, by arguing that it was in fact racist to vote for the Voice and that people had to vote ‘No’ to prevent Australia entrenching racism in its Constitution.
Next, the Conservative No campaign contended that the amendment was ‘legally risky’. It raised fears that the High Court of Australia would interpret it as requiring prior consultation of the Voice in relation to all government decisions, resulting in the courts being blocked up with judicial review applications if the Voice did not get its way. It claimed that Government would become dysfunctional as all decision-making became bogged down in litigation.
A third argument was that there was insufficient detail about how the Voice would be comprised or operate, as this would be left to Parliament to determine. This argument fed into a distrust of politicians and Parliament to resolve the problems of Indigenous disadvantage and concerns that the Voice would either be too weak or too strong, and would be dominated by Indigenous ‘elites’ rather than people from remote areas who were better acquainted with the most serious problems facing Aboriginal and Torres Strait Islander peoples.
There were, of course, rational counter-arguments to all these arguments, but they required greater explanation and were not easily distilled to sound-bites. It takes more than 10 seconds to outline the concept of substantive equality and explain that it does not require treating everyone exactly the same regardless of their different characteristics and circumstances. Moreover, such explanations required a degree of trust in those giving them, and trust was singularly lacking in the social media din, especially during a period marked by cost of living stress.
The absence of detail about the architecture of the Voice also fuelled a narrative of secrecy and Trojan Horses, which was particularly attractive to a constituency of the ‘No’ campaign that included sovereign citizens, those who feel oppressed or neglected by Government and those who are vulnerable to conspiracy theories.
Social media effectively spread many conspiracy theories, including that the Voice referendum: would result in people losing their homes and backyards to Indigenous claimants; was a plot by the United Nations to take control of Australia; would give effect to a coup to put in place an un-elected Indigenous monarchy and apartheid; would establish a separate tier of Indigenous government; or would turn Indigenous Australians into British subjects and strip them of their land rights. Trying to put out the spot-fires of these more extreme claims had the undesirable effect of shining attention on them, fuelling the general impression that the proposal was ‘risky’ and best abandoned.
There were a number of other factors that influenced the outcome.
The first was compulsory voting which applies both to elections and referendums in Australia. On the one hand, this ensures the outcome is a more accurate reflection of the views of voters, and bolsters the credibility and legitimacy of the vote. The outcome of a referendum does not just reflect the views of those on the extremes and those who are more successful at ‘getting out the vote’. However, compulsory voting does also mean that those who have no view and are not prepared to educate themselves to form a view, still vote. In an election, they may choose to vote informally, but in a referendum they tend to vote for the status quo, which means they vote ‘No’. Political scientists have estimated that compulsory voting adds 7% of voters to the ‘No’ cause, raising the bar for ‘Yes’ to win.
Treaty, sovereignty and other Indigenous policy issues
A significant portion of the campaign was spent debating other matters, such as the negotiation of treaties, the establishment of a truth-telling process, whether reparations would be payable for past confiscation of land, the status and meaning of Indigenous sovereignty and whether existing rights to land would be affected. There was confusion within the community about whether they were voting on these matters, as opposed to the actual constitutional amendment.
Degree of Indigenous support
The degree of support for the referendum proposal amongst Indigenous Australians was also an important and contested factor. The Voice proposal had been a key recommendation of a constitutional convention of Indigenous Australians, which was the culmination of a consultation by way of dialogues held across the country in Indigenous communities. It therefore had broad support amongst disparate Indigenous groups, as well as by Indigenous organisations, such as Land Councils and peak bodies.
However, the ‘No’ campaign was led by three Indigenous Australians – Senator Lidia Thorpe on the progressive left side and Senator Jacinta Nampijinpa Price and Nyunggai Warren Mundine on the conservative side. Indigenous leadership on the ‘No’ side gave cover for arguments that might otherwise have been regarded as socially unacceptable. Voters also asked why they should support a measure to recognise and aid Indigenous Australians, if Indigenous Australians were not themselves agreed that it would be desirable and effective.
The risk and permanency of change
The difficulty of changing the Constitution is to an extent self-fulfilling. The perceived risk of change is enhanced because voters know that once something is enshrined in the Constitution it cannot be changed by legislation or corrected if it has unintended effects or is interpreted by the courts in an unexpected fashion. The only way to alter it is to hold another referendum to change the Constitution, which is unlikely to succeed. The perceived permanence of change makes voters reluctant to support it, especially if they feel ill-equipped to understand the proposal or assess the consequences of change. Polling research by DemosAU – ‘Voice to Parliament Research – What Drove the No Victory?’ – concluded that the ‘determining factor in this referendum [was] fear of constitutional change’.
Both the Yes and No campaigns engaged in advertising – with the No campaign concentrating on online advertising through social media and focusing on the States with smaller populations (as an Australian referendum requires support in a majority of States to succeed). It seized the narrative early in the campaign, leaving the Yes campaign struggling to turn it around later in the campaign.
According to the Resolve Political Monitor on the Voice Referendum Poll in October 2023, the Yes campaign was more effective in swinging votes through door-knocking and direct conversations with voters, but less effective with its advertising.
The Result and its demographic spread
The result was roughly a 40-60 split, with 39.94% of valid votes for ‘Yes’ and 60.06% of valid votes for ‘No’. Voting was compulsory, with a turnout of 89.95% (almost the same as for the 2022 election) and a very low informal vote of less than 1%. The referendum failed to obtain a majority in any State, with only the Australian Capital Territory achieving a majority Yes vote of 61.29% approving the constitutional change.
Overall, electors in the 18-34 age group voted in favour of the referendum, with electors aged 35 and over tending to vote ‘No’, with the likelihood of a No vote increasing with a person’s age. The electorates that voted Yes were almost exclusively inner metropolitan seats or seats in large regional cities. The No vote dominated in outer suburban, regional and remote areas. For most of the campaign, females were more likely to say they would vote Yes than males, but this began to even-up towards the end of the campaign, so that sex was no longer a significant distinction.
Electorates with higher levels of affluence were more likely to vote Yes, but the biggest predictor of voting outcomes was level of education, with electorates with high numbers of university graduates voting Yes, while those dominated by people with high school or trade qualifications tended to vote No. All of the nine electorates in the country where more than half of the population has at least a bachelor’s degree voted Yes.
While there was ongoing debate about how Indigenous electors voted in the referendum, with much of the analysis falling into ‘ecological fallacy’, there was some evidence to show that at a polling booth level, those booths with predominantly Indigenous electors voted Yes.
Prospects for future reform
The defeat of the referendum makes it far less likely that future referendums will be held. The current federal Government had previously floated the prospect of a republic referendum as a ‘second term’ possibility. But the defeat of the relatively simple Voice referendum suggests that a far more complex republic referendum in a politically charged environment would have little to no chance of success. So the Constitution remains frozen in cryostasis, awaiting a political thaw in a future age.