Skye Predavec
Researcher

A reader of the blog has asked: how many more seats would the Coalition lose if an election were held today (based on the latest Newspoll)?

On Sunday night, the latest Newspoll showed a new record low for the Liberal–National  Coalition, but – if repeated in an election – how would that shape up in terms of seats? And is that even the right question?

The poll reported Labor has 36% of the primary vote, the Coalition 24%, Greens 11%, One Nation 15%, and other candidates (including independents) 14% collectively. On two-party preferred, Labor leads 57% to 43% (up just shy of two points from the election in May).

The traditional model for Australian elections is the Mackerras Pendulum, which places all seats on a list according to the margin they’re held by. If the pendulum swings in Labor’s direction, marginal Coalition seats fall, but if it swings the other way, Labor loses its marginal seats.

Assuming the swings in this poll apply to all seats:

·       Labor would win 97 seats, up three from the election

·       The Coalition would win 40 seats, down three

But this approach is limited, to say the least.

Australia Institute research has found that the Pendulum only correctly predicts the winner in a third of seats. The national vote total does not determine the results in individual seats, and that’s further complicated by the rise of non-major parties and independents.

To the right candidate, with the right campaign, no seat is safe – regardless of the national numbers.

Take the Greens in the last election, for example. Despite recording a swing of only 0.05% against them, the party lost three of its four seats to Labor because of drastic changes in where their vote came from. So, while the latest Newspoll poll implies the party would lose its remaining seat of Ryan to Labor, that’s far from certain.

Similarly, applying the Newspoll primary vote swing to One Nation to this year’s election results suggests they will pick up one seat – Wright – and independent Zoe Daniel would re-take Goldstein. But just as the Pendulum could not predict the independent surge in 2022, it cannot predict how voters will respond to candidates who have not yet decided to run, let alone announced their candidacy. Trying to project the outcome of an election in three years’ time from a single poll is a fraught exercise at best, especially when it misses the real story: Labor and the Coalition are together polling just 60% of the primary vote, a record low for the major parties and down 6 percentage points from the election.