During the month of May there were Australian by-elections on three successive Saturdays. There was a swing in votes to the right in all three cases and the Greens performed badly in all three. Here’s my very tentative prediction for the May 2028 general election.
During the month of May there were Australian by-elections on three successive Saturdays. On 2 May the election was in the Melbourne outer-metropolitan seat of Nepean. On 9 May it was for the federal rural seat of Farrer (NSW) and on 16 May it was in the Brisbane inner-metropolitan seat of Stafford. Of the federal parliamentary parties only the Liberal Party and the Greens contested all three seats with the Liberal Party performing very well in Melbourne and in Brisbane where the Liberal Party is known as the LNP. Nepean and Stafford were for the state Legislative Assembly while Farrer was for the federal House of Representatives.
There was a swing in votes to the right in all three cases and the Greens performed badly in all three. Pauline Hanson’s One Nation Party was done like a dinner in Nepean and – very wisely – avoided a similar humiliation in Stafford by not contesting the seat. However, the win for One Nation in Farrer has sparked a new round of pundit opinion expressing the nonsensical view that the Liberal Party is suffering from an existential crisis.
Let me explain quite how well the Liberal Party has performed in Nepean and in Stafford. Nepean was last won by Labor in November 2018 when the winner, Chris Brayne, secured 50.9% of the two-party preferred vote. In November 2022 he was defeaed by Liberal Sam Groth who secured a swing of 7.2%, so won with 56.3%. After the cynical and selfish resignation of Groth, the new Liberal member, Anthony Marsh, has now won with 63.3% (in competition with a leftish independent) so he has achieved a further swing of an even 7%. The total swing to Liberal, therefore, has been 14.2% since 2018.
Stafford was last won by the LNP in March 2012 when the then new LNP member, Dr Chris Davis, secured an even 57% of the two-party preferred vote in the landslide general election win for the LNP by Campbell Newman. It is a natural Labor seat and won by Labor at every election since Newman’s day. However, on the 2026 by-election figures (and allowing for boundary changes to be implemented for the October 2028 election) it now becomes Labor’s fourth most marginal seat in Queensland, the four being Beenleigh, Pine Rivers, Springwood and Stafford. Labor-held Gaven becomes notionally LNP on the new boundaries.
At the October 2024 Queensland state election, the then Labor member, James (Jimmy) Sullivan, secured 55.3% and the LNP candidate, Fiona Hammond, won 44.7%. The by-election figures are not final but it seems to be 51.4 for the new Labor member, Luke Richmond, and 48.6 for Hammond who is contesting for the LNP again. So the swing will be about 4% to the LNP.
New governments tend to perform well in by-elections for Opposition seats. For example, Albanese Labor took the federal seat of Aston from the Liberal Party at that seat’s by-election in April 2023. The problem with making this a general rule is that almost every such case has been when an Opposition member cynically and selfishly resigns, of which Aston and Nepean are cases. Stafford was caused by Sullivan’s death, which leads me to make the counter factual speculation that if it had been caused by the typical selfish and cynical resignation the LNP would have won Stafford – but Labor has held on to this now highly marginal seat.
All in all I think the Liberal Party is looking pretty good in both Victoria and Queensland. I have never doubted that the Crisafulli Government will win a second term in October 2028 and, as I foreshadowed in my early February 2025 article “Setback looming for Victorian Labor” I am now confidently predicting that the November 2026 Victorian state election will see the installation of a Liberal- National coalition government with an absolute majority in the Legislative Assembly.
Sandwiched in betweeen Nepean and Stafford came the Farrer by-election, the result of which has caused all this commentary about the future of the Liberal Party. Certainly, when a party’s share of the primary vote falls from 43.4% in May 2025 to 12.4% in May 2026, it looks very bad for that party. However, I argue that it is not as bad as it looks. I remind readers that in my article dated 16 February titled “Disruptor David Littleproud enjoys another win” my immediate prediction was that the Liberal Party would lose the seat – so there was no surprise in the result because commentators have known all along that Farrer is not a natural Liberal seat. The consistently high Liberal vote this century was the consequence of the coalition agreement, meaning that most of the wins for Sussan Ley were on National Party votes.
I refined my predictions in my later article dated 16 April “Farrer by-election prediction: One Nation to Win”. However, even there I made a trivial error. My prediction for the order of primary votes won was first David Farley (One Nation), second Michelle Milthorpe (Independent), third Raissa Butkowski (Liberal), fourth Brad Robertson (Nationals) and fifth Richard Hendrie (Greens), but in the end the Greens did not come fifth because the Legalise Cannabis Party candidate, Aimee Pearson, won 2,357 primary votes, while Hendrie of the Greens won 2,346 votes. (At the time of writing the AEC website does not show the detailed preference distributions. It could be that after the distribution of preferences from six low scoring candidates Hendrie will finish fifth. However, this detail illustrates quite how badly the Greens have performed at these by-elections.)
The final two-candidate statistics in May 2025 were 57,916 votes for Sussan Ley (56.19%) and 45,147 votes for Milthorpe (43.81%). For the by-election they are 58,294 for Farley (57.55%) and 42,992 for Milthorpe (42.45%). The swing to the right, therefore, has been 1.36%. Ley’s margin over Milthorpe was 12,769 votes. Farley’s margin is 15,302 votes.
Following the Farrer by-election, a member of my fan club wrote an email to me with these questions: “Do you think this is a sign of One Nation becoming a new ‘third force’ in Australian politics and potentially supplanting the Coalition by taking all its rural and regional seats? Or do you think it will implode and fracture as time goes by?”
To give an answer, I begin by stating the numbers of seats in the present House of Representatives. They are: Labor 94, Liberal 27, National 14, “teal” independents six, One Nation two, Greens one and all the rest combined six, for a total of 150. My very tentative prediction for the May 2028 general election is Labor 78, Liberal 40, National 14, One Nation four, “teal” independents three, Greens two and all the rest combined nine. That makes a total of 150 again. The four One Nation seats will all be very rural in character.
Commentators would then describe the results as unexpectedly poor for One Nation, which party would proceed to break up. David Farley would then go on winning Farrer as an independent – but Farrer will eventually revert to being a Nationals seat.
I realise that much of the above runs counter to the opinion polls for May. My response is to say that mid-term polls reflect the kinds of answers people are prepared to give on the spot to questions uninvited and about which they have not had time to think.
