There’s been a lot of discussion in the last few days about the merit of Australia’s electoral system. In response to some conservative attacks on our preferential system, sometimes implicitly or explicitly suggesting a first-past-the-post system would be somehow more legitimate.
In response to that, Kevin Bonham has a good piece in the Guardian dismantling the pro-first-past-the-post arguments. In Crikey, they set up a debate between two pieces, one by William Bowe defending the electoral system, and one from Robert Lechte arguing that the electoral system does in fact “suck”.
While I am very much not a supporter of first-past-the-post. I think there is a danger in defending our system that we ignore the problems. While Lechte did make a case for proportional representation, I disagree with a lot of the arguments, let alone his irrelevant side-swipes on a variety of issues.
So I wanted to make my case here on one specific but very important element of the electoral system: the use of single-member electorates to elect the House of Representatives, as opposed to some system of multi-member electorates using proportional representation.
This is not my only issue with the electoral system, but I should say that generally the way that Australian elections are administered is excellent. The Australian Electoral Commission does an excellent job, is generally well funded and is able to be well trusted and independent. We also make it easy for voters to get to the polls and cast a vote, and our system of redrawing electoral boundaries is generally very good.
Activists have been campaigning for proportional representation in Australia for well over one hundred years, dating back to Catherine Spence who campaigned for “effective voting” in the 1890s, and eventually was Australia’s first woman candidate when she contested South Australia’s election for delegates to the 1897-98 constitutional convention.
A lot of arguments have been regular and constant throughout that century, but I’m not going to spend a lot of time on those.
Yes, a majoritarian electoral system leads to governments winning majorities when they don’t win a majority of the vote, and that is increasingly true now, with the current Labor government polling less than 35% of the primary vote. I think this remains true even when voters have the option of marking a further preference. It is still the case that barely one third of the country most preferred Labor in government. While they won the two-party-preferred vote with at least 54%, all that tells us is that voters preferred Labor to the Coalition (another reason why the pro-FPTP advocates are full of it), not that they preferred a Labor majority government to any other government shape.
And the majoritarian system means that the electoral contest becomes focused on a minority of electorates while most remain safe, thus creating incentives to focus attention on some areas while ignoring others. Even though once-safe seats have now become marginal, on average seats are just as safe as they once were.
The median electorate has a two-candidate-preferred margin of 9% , which is the highest we’ve seen since 2004. There are more seats where a seat is marginal on the three-candidate-preferred margin, but not that many.
But as the primary vote for independents and minor parties has continued to climb, I would argue that a lot of the problems with the majoritarian system have become worse, and things that were once key selling points of the majoritarian system no longer apply.
For a start, disproportionality in parliament is off the charts. The Gallagher least squares index measures the gap between the vote share and the seat share of parties. Michael Gallagher has calculated this statistic from 1946 to 2022. He treats the Coalition parties separately, something I think you could justify until the Liberal National Party merger in 2010 but harder to do now. So I have calculated my own statistic treating the Coalition as one party, back to 1984. The two numbers move closely together, but not identically.
Both statistics reached a record high in 2022, but I estimate the number is much much higher in 2022, reaching a score of 22. In contrast, the 2024 Tasmanian election, conducted using the proportional Hare-Clark system, produced a score of 3.5.
While preferences are a helpful tool in broadening the popular mandate of MPs, the fact remains that the first preference is a clear indicator of who a voter supports. And it is becoming more and more removed from the results. Another way to show this is to measure what proportion of voters gave a first preference to a winning candidate.
This number dropped below 50% for the first time in 1990, but dropped precipitously in 2022, and has dropped slightly further again in 2025. Just 41.3% of voters gave a first preference to the winner.
This is not an argument for first-past-the-post, which would slightly improve this statistic, but instead for using a system where more than one person can be elected, so a number of main voting blocs can have a local representative. Again to contrast with Tasmania, over 60% of voters gave their first preference to an individual candidate who won, and 88% voted for a party that won a seat in their electorate.
The preferential voting system is particularly bad at handling three-cornered contests. Once there is some doubt about who will be in the top two, it creates strategic voting incentives for voters, and creates the potential for non-monotonic outcomes.
In Macnamara in 2022, a voter who voted 1 Liberal but is most concerned about defeating the Greens almost helped the Greens win the seat by pushing Labor below the Liberal. In Fannie Bay in 2024, Labor won the two-party-preferred vote but the Greens lost the two-candidate-preferred vote to the Country Liberal Party. If some Greens voters had switched to Labor, they would have ensured the CLP did not win.
It seems likely that some independent seats in 2025 will be in this category. We won’t get a Liberal vs Independent 2CP count in Grey or Monash, but once we have final 3CP figures we’ll be able to make estimates, and I suspect those independents may have won such a count.
I think we can mostly agree that our electoral system, or something like it, is the best way to elect a single person to parliament. But when you elect a lot of people, the same narrow and arbitrary results in numerous seats can produce disproportional outcomes. Right now Labor is benefiting, but it won’t always be like that. In 2022, we saw two localised results where a minor political forces won multiple seats in the same area with a relatively small share of the vote. The Greens won three seats in inner Brisbane on about a third of the vote, while the teals won three out of four seats on the north shore of Sydney on about a quarter of the vote.
Some people criticised me for describing the result where the Greens won Ryan and lost Griffith precisely because Ryan is a more conservative seat as “perverse”, but I had also suggested the Greens sweep of inner Brisbane in 2022 was strange in a 2023 paper.
This also makes counting and analysing the results more complex, but that’s not really the point here. It certainly makes things more interesting. But it’s not great for the voters, and the ultimate outcome is often quite volatile and arbitrary.
All of this means that the relationship between how people vote and who gets elected has become weaker, and less logical. I don’t think this problem is going to go away, and we should acknowledge it.
My solution? A system of proportional representation, likely with multi-member districts. I would prefer 5-member districts but even three-member districts would significantly improve proportionality and reduce the arbitrary impact of those three-cornered contests.
It is very easy in punditry to be very focused on the race in front of you, on calculating who will win and marvelling about the unsual and interesting calculations needed to know who would win. But I think it’s important to also zoom out and judge our electoral system normatively, not just treat it as a fact that cannot be changed. The way our system is working now is weird, and isn’t producing great outcomes.