I’ve already written extensively on the One Nation ‘orange wave’ and particularly on the paradigm-shattering state by-election in WA last weekend. One part I skipped, though, was that, yet again, the chin-stroking pollsters and the bookies were dead wrong. Why they were wrong is the real story.
As One Nation rockets up the polls, the finger-wagging narrative from both heads of the uniparty is that a primary vote for ON is ‘wasted’, because the establishment shills will win on preferences anyway. It may seem plausible enough, when you consider that the Albanese government garnered one of the lowest primary votes in Labor’s history, yet won a whopping 94 seats on preferences.
Is the narrative about One Nation and the two-party preferred result true, though?
The real-world acid test in Secret Harbour showed that preference flows will likely be much better for One Nation than the establishment wants you to believe.
Every single one of these polls, any of the polls that have a two-party preferred calculation involving One Nation, every single one of them is wrong. And not by a little bit, wrong by a lot.
The reason is not mysterious. Pollsters have to hang their two-candidate models on something. Like generals fighting the last war, the something they hang them on is last election’s preference flows, with maybe an ‘We Just Reckon’ effort at updating them.
That would be fine if the electorate were a statue. It isn’t. It’s a living, breathing, voting thing. One Nation has, since the last election, grown from just over six per cent of the vote, to at least in the mid-20s. Most polls now put them ahead of both Labor and Liberal, federally. Yet, we’re supposed to believe that a party can quadruple its primary vote, but preference flows will remain locked into 2025. Given that most of the voters who’ve flocked to One Nation have done so fleeing from the coalition, it seems highly unlikely they’d direct their preferences to Labor or the Greens. Of the shrinking minority who are sticking with the Liberals or Nationals, it seems even less likely that they’d put One Nation last.
Governments rise and fall on preference discipline. In fact […] just correcting that turns a One Nation loss into a win just by using the most up-to-date numbers for this thing called preference discipline […]
What’s been true for a long time in Australian politics is that the left of politics tends to have very good preference discipline. Meaning that votes for, let’s say, the Greens or voters for the Greens, they will preference Labor pretty much 100 per cent of the time over One Nation […] all these voters that are identifiably on the left of politics, they tend to do a very good job of listing all of the left-wing parties in their preferences before they include anybody who isn’t left-wing. Essentially, the votes in the left-wing basket stay in the left-wing basket. But the same hasn’t historically been true on the right of politics.
Secret Harbour changed all that.
One Nation’s Luke Herdegen took 39 per cent of the primary in a seat Labor had treated as family silver. After preferences he finished about 57–43. The interesting number is not the final margin. It is where the extras came from. Add the identifiably right-of-centre primaries that were not One Nation – Liberals, Libertarians, Australian Christians, Shooters – and you get a pool of a bit over 5,300 votes. Of those, One Nation collected about 4,700. That is a right-of-centre flow of 88.8 per cent. By contrast, in Farrer just a few months earlier, only 62 per cent of Liberal leftovers stayed on the right. And even that was in a very singular contest, Labor having not bothered to stand, leaving a locally popular independent as the left’s vehicle.
Cue the narrative managers. A numbers-fiddler of the ‘I have no party loyalties but I quote Crikey and endlessly rag on the right’ variety announced that One Nation’s preference flow was a soggy 49 per cent. Technically he can add the figures that way. He is using the wrong denominator. He is counting every preference that moved at all, including the left’s, which were always going to Labor. That is a bit like judging a rugby team’s score by including the tries that went to the other team.
The number that matters for a Labor-versus-One Nation two-party contest is simple: when a right-of-centre voter is forced to choose, do they stay right? In Secret Harbour, almost nine in 10 did.
Apply that to the published polls and the ‘Labor wins easy on preferences’ narrative starts to look like advertising. Roy Morgan and Redbridge have been effectively leaking seven points from a 27-point coalition-plus-others pool across to Labor. That is how they manufacture a 55–45 or 52–48 Labor two-party from primaries that no longer support it. Lift the assumption to Secret Harbour’s 89 per cent and those same polls flip, or come within a sneeze. A comfortable Labor 2PP becomes a One Nation win.
Are the pollsters lying? Probably not. They know last election’s flows are stale. They also know that anything else is a guess, and guesses get you laughed at when they miss. So they keep publishing 2PP tables they privately treat as ornamental, while the uniparty waves them at nervous voters: don’t waste it, darling, they’ll never get there on preferences.
Secret Harbour says otherwise. The wasted-vote sermon was always a campaign tool. The data from Perth’s southern suburbs just took the pulpit away.
