One thing you can’t accuse the pollsters of doing during the referendum debate is “herding,” that sometimes alleged tendency of these outfits to twiddle their results to avoid standing out from the crowd. The most recent findings, at least at the time of writing, range from Newspoll’s 63–37 for the No forces to Roy Morgan’s much more modest 54–46.
You never see anything near that kind of variance in ordinary voting-intention polls, though that’s at least partly because the pollsters are posing quite different questions from each other this time. Whose is the best? Referendum polls are a more imprecise science than election ones, so let’s talk about it after the results are in.
But no pollster has Yes ahead, alas, or even approaching it.
At last year’s federal election, people who followed the count from 6pm Australian Eastern Daylight Time experienced an early burst of déjà vu: was this 2019 all over again? The first numbers came predominantly from Tasmania, which happened to be the only state or territory to swing to the Morrison government. By around 7.30 the direction and approximate size of the national swing was clear, and it was an altogether different direction.
There’s a good chance the Apple Isle will again be the renegade on Saturday night, registering the biggest Yes of all states and perhaps generating some excitement for a while. But with only two piles for the ballot paper to be sorted into (plus one for informals) the Australian Electoral Commission’s counting should be faster than at a general election. Even under the Morgan scenario the overall result should be clear by 7pm.
Yet the telecast is scheduled to last until 10pm. What on earth will they talk about?
A referendum passes if it receives a majority of national votes and a majority of votes in a majority (at least four) of states. House of Representatives electorates have nothing to do with it, though they will be of academic interest; according to Antony Green the ABC website will publish all electorate results live online and discuss ones of interest on TV.
Tasmania, that wildest card, has a proud anti-Canberra — or maybe just anti-mainland — tradition of returning relatively high No votes in referendums, but this time the polls, albeit with small sub-samples, pretty consistently show it as the most supportive.
Not coincidentally, Tassie has the only state Liberal government in the country, and that government is supporting the Voice. And the high-profile federal Liberal member for Bass, Bridget Archer, a long-time irritant to her party, is energetically advocating Yes. Voters love major-party mavericks, and not just in their own electorates. But might the Tasmanian government’s meltdown over the past fortnight dampen the effect?
Of the states and territories up on the mainland, the city-territory of the ACT will record the highest Yes and Victoria can be expected to come second — unless South Australia or New South Wales does. It would be very surprising if Queensland and Western Australia don’t bring up the rear. But the rest is rather unpredictable.
Counting in South Australia and the Northern Territory won’t start until 6:30pm AEDT and Western Australia 9pm.
As a broad rule, the “big” states of New South Wales and Victoria, plus South Australia — second-least populous but displaying less of an outlier mentality than Queensland and Western Australia — have been relatively supportive of referendums. We saw it, for example, with both questions at the last set in 1999.
(In 1988, in a very rare break from tradition, Queensland recorded the biggest Yes votes for all four because one of them, for “free and fair elections,” was expected to apply particularly to the electoral malapportionment engineered by the Bjelke-Petersen government.)
Pollsters are broadly predicting these state trends to continue tomorrow. And yet… New South Wales came stone-cold last of all states and territories in the 2017 marriage equality survey. In crude terms, this was due largely to socially conservative religious white people outside the capital and many socially conservative religious people of colour in Sydney, particularly in the west. While the former can be expected to vote heavily No to the Voice, some pollsters have found people from non-English-speaking backgrounds/immigrants more likely to support Yes nationally than the rest.
That last expected dynamic seems counterintuitive. So it will be worth tracking seats in Sydney’s west, and consequently New South Wales as a whole, on the night.
The “inner-urban” professionals–heavy, high-income electorates, including the ones with teal MPs, can be expected to lean Yes. Not so the middle- to high-income, low-NESB ones such as (in Sydney) Lindsay, Hughes and Cook.
Does it go without saying that young people will be much more supportive of the Voice than old ones? That can’t be determined by the voting results, only by surveys. For gender, too, we must rely on opinion polls. For some other cohorts, where census data indicates they constitute a big majority in certain polling places, analysis can provide clues. (Ideally such analysis would adjust for aggregate state or territory votes.)
This time the topic of the referendum adds another geographic dimension. Attention will focus heavily on Indigenous booths, particularly in the Northern Territory, which has more First Nations voters than any other jurisdiction.
The electorate of Lingiari, although only 40 per cent Indigenous (with a lower proportion among voters, given a relatively low median age and a lower, if recently improved, enrolment rate), will come in for outsized attention, but the seat’s aggregate vote won’t tell us much. Turnout in Lingiari is another matter, however, and is (at the risk of tempting the ecological fallacy gods) an okay indicator of, if not the exact level, then the seriousness of low turnout among First Nations voters. In June last year the new prime minister made a big fuss about its low 65.8 per cent.
“That was part of the former government’s design. It wasn’t by accident,” thundered Anthony Albanese, “and they should be held to account for it!” He had a point — the Coalition wasn’t much interested in raising Indigenous participation — but only a small one, because the low Indigenous turnout there reflected a range of factors.
The Australian Electoral Commission has in the last year pulled out many stops to bring estimated Indigenous enrolment up to near the national figure, presumably at least partly by selectively relaxing the rules of its direct enrolment and update program. That’s excellent, but it will — all else being equal — make the turnout worse; after all, the figure will show voting as a percentage of the roll, not of estimated eligible people.
But all else won’t be equal for this referendum. We can expect a higher degree of interest from those voters, and the AEC is pulling out many stops to get their ballots. No doubt others on the ground are also encouraging people to register their vote.
When the counting is done, Lingiari’s turnout will attract attention. It might or might not be higher than at the 2022 election. But the more comprehensive Indigenous electoral roll will come back to bite Anthony Albanese at the next general election.
Probably a dozen or two of our 151 electorates will vote Yes, and maybe, perhaps, one state, Tasmania. The expected thrashing is consistent with your correspondent’s long-expressed deterministic mindset. Midterm Labor government referendums get slaughtered, partly because Liberal opposition is all but inevitable. The content of the proposal only affects the size of the defeat. And this would also apply to merely “symbolic” recognition: the Liberals in opposition would have opposed that too, regardless of what they say (or believe) now.
So expect Saturday night to degenerate into indulgent hot takes as self-identified campaign “old hands” compete to tell the tallest tales, and how they would have run Yes instead. The endless analyses, the gloating dressed up as advice will not be for the faint-hearted. And it won’t stop then but will last for weeks, no doubt joined by intra-Yes sniping.
What does it say about Australia? To this observer, the result will tell us nothing we didn’t already know.
So, no, the No campaign wasn’t brilliant, and the Yes one wasn’t that awful. The devastation about to engulf people on the affirmative side, particularly those for whom this has been a decade-long project, might not be leavened by this knowledge, but this result was a fait accompli once the decision was made to hold the vote apart from an election.
Amid the celebrations, Peter Dutton might even believe this will save his leadership. Perhaps he should have a word with his mentor John Howard, who likely reckoned the same in September 1988 but was out of a job eight months later. •