Swings on the pendulum

Three weeks ago, when we looked at some of the Australian election numbers, I promised a further post when we had complete two-party-preferred figures. They are still not quite complete, because the electoral commission has not conducted a two-party count in Bradfield – evidently it doesn’t want to disturb the ballot papers in case there’s a legal challenge to the 26-vote margin by which Teal Nicolette Boele was declared elected. (Ben Raue had an interesting if rather technical post about this earlier in the week.)

But one seat isn’t going to make much difference, so with the rest all decided we can say some things about how the election turned out in terms of the choice between Labor and the Coalition (now back together after a brief separation).

The first thing to say is that for an increasing number of voters, that’s not the primary choice that they made. A record 33.6% gave their first preference vote to someone other than the major parties, and in 35 of the 150 seats the final contest was not between Labor and the Coalition. You can see the list here: ten of them involved the Greens, three the far right, and the rest all independents (including Centre Alliance in that category).

Nonetheless, it remains for now fundamentally a two-party system, and the two-party-preferred vote is still an effective way to study it. On that basis Labor won with 55.3%, its best result since 1943 and a swing of 3.2% from 20192022. This is unprecedented; no previous first-term government had recorded a swing in its favor since the establishment of the two-party system.

Since Labor’s primary vote was 34.6%, it picked up in aggregate just over twenty points from the preferences of the non-major-party candidates, while the Coalition picked up about 13 points. That proportion is almost exactly the same as last time.

Labor won the two-party-preferred vote in exactly two-thirds of the seats, 100 out of 150, up 16 from 2022 (assuming it doesn’t win it in Bradfield), although one of those (Aston) it had already picked up in a by-election. Since Labor finished with 94 seats, that means it won the two-party-preferred vote in six of the 13 seats won by crossbench MPs, while the Coalition won it in the other seven.* In other words, Labor’s underlying lead over the Coalition is much the same as its actual lead in seats.

Of those seven crossbenchers who sit for what in two-party terms are Coalition seats, five – the members for Bradfield, Curtin, Indi, Kooyong and Mackellar – are Teals or Teal-aligned. But another two of the canonically Teal seats, Warringah and Wentworth, now have an underlying Labor majority. That doesn’t necessarily mean Labor would actually win them if there were no Teal candidate, but given that they were once among the safest of Liberal seats it’s still a remarkable development.

All but 27 of the seats (plus Bradfield) swung to Labor, and most of the pro-Coalition swings were quite small, although Bendigo swung by a remarkable 10.7% (note that the AEC’s table doesn’t report that because the identity of the Coalition candidate changed from Liberal to National). Solomon in the Northern Territory swung 7.1% and another four seats (three of them in Western Australian) recorded more than 5%.

Biggest of the pro-Labor swings was Braddon in Tasmania with 15.2%, leaping from fairly safe Liberal to fairly safe Labor without pausing at marginal. Fowler (12.3%) and Lyons (10.7%) also hit double figures and another 28 seats had swings that topped 6%. The median swing was 3.5% to Labor, just slightly above its average swing. About 60% of the seats were within three points either side of that figure.

And what about the pendulum? Going by the pre-election pendulum, a uniform swing of 3.2% would have delivered Labor a notional gain of 13 seats, so its actual gain of 15 (not counting Aston) is a little better but very close. It fell short in Canning, Casey, Longman and Monash, but instead picked up Bonner, Braddon, Forde, Hughes, Leichhardt and Petrie. Swings are never uniform, but deviations roughly cancel out.

To win a notional or underlying majority next time (and assuming no relevant boundary changes), the Coalition will need to pick up 26 seats, which would require a uniform swing of 6.9% (Paterson is the tipping-point seat). But even that daunting task will not be enough unless, in addition to improving its position against Labor, it can also win back seats from the crossbench, particularly the Teals – or else win their backing in a hung parliament.

The more the crossbench expands the less useful the pendulum becomes, but for now it provides us a snapshot of the depth of the Coalition’s problem.

.

* Strictly speaking that doesn’t follow; it would be possible for a party to win a seat while losing the two-party-preferred vote, if the other major party came third and favored it with its preferences. But that would be very unusual and did not in fact happen anywhere.