UPDATE: This post was published on 19 May before the official distribution of preferences. The produced a finally tally where Liberal Gisele Kapterian led Independent Nicolette Boele by 8 votes. A full re-count of all first preference votes, a re-check of all informal votes, and a new distribution of preferences will now take place. Apart from this additional detail updating the final count, the rest of the post remains correct.
Bradfield, once the Liberal Party’s safest metropolitan seat, has turned out to be an absolute squeaker at the 2025 election.
After all primary votes have been officially tallied, the indicative preference count has Independent Nicolette Boele defeating Liberal candidate Gisele Kapterian by 39 votes.
The narrowing of the result in the last week has been a surprise. Independents do not traditionally do well on Postal, Absent and Pre-Poll Declaration votes. And if you look at the primary votes, Ms Boele has not done well.
But the Greens and Labor HAVE been doing well on Absent and Pre-Poll Declaration votes, and it is these that have boosted Ms Boele’s after preferences count in her second attempt to win Bradfield.
So where does it go from here. Let me try and explain.
Nicolette Boele may lead by 39 votes after preferences, but this is obtained from what’s called an indicative preference count.
Indicative counts were introduced in 1993 to allow candidates, the media and other observers to know how the count’s going ahead of the formal distribution of preferences.
Two candidates are nominated beforehand as the final pairing, and all ballot papers for other candidates are examined to determine which final candidate will receive each ballot paper’s preference.
But the indicative count does not determine the winner. An official and full distribution of preferences needs to be undertaken.
This is the count you were taught about at school where the lowest polling candidate is excluded, their ballot papers examined to determine the next preference, and ballot papers transferred to a continuing candidates.
This process of exclusion and transfer continues until only two candidates remain.
It is this full distribution which will now be undertaken. With scrutineers watching the count closely, the distribution will take about a week. Once completed, an official winner and margin of victory will be declared.
If the final margin is over 100 votes and the loser chooses not to ask for a re-count, the writ will be returned with the name of the winning candidate.
If the winning margin is under 100 votes the AEC does an automatic re-count. If it is above 100 and the losing candidate has grounds to request a re-count, a re-count may also be undertaken.
The re-count will re-examine the formality of every formal and informal ballot paper to determine a new tally of first preferences by candidate. Then a second distribution of preferences will take place.
Only one re-count is allowed. The winner will be declared elected after the re-count and the writ returned with the name of the duly elected candidate. The elected member can be sworn in and take their seat in the House of Representatives. They can take their seat even if there is a court challenge.
If the re-count produces a tie the writ cannot be returned with a named winning candidate, it is referred to the Court of Disputed Returns which would quickly order a supplementary election.
But the result can be challenged in the Court of Disputed Returns. The Chief Justice of the High Court would nominate a High Court or Federal Court judge to sit as the Court of Disputed Returns.
In law the case is to argue that the person named on the returned writ should not have been declared elected. The Court process is to establish errors in the conduct of the election that the petitioner argues should result in the election being voided.
An issue like corruption can lead to an immediate voiding of the result.
If the problem is with the conduct of the election, the issue is whether enough errors can be found to put the result in doubt. Just finding errors isn’t enough to void an election. It must be proved there are enough errors to bring the result into doubt.
In theory the Court can change the writ to name a new elected candidate, but this is unusual.
More likely the Court will either let the result stand or order a new election.
But will the losing candidate challenge? If a new election were the outcome, would the losing candidate be able to win the re-election?
It is an important political consideration before choosing to challenge.
After the 1996 Federal election, the Labor Party challenged the eligibility of Jackie Kelly as elected member for Lindsay on constitutional grounds. It won and forced a re-election where Kelly was elected with a 5% swing in her favour.
At the 1995 Queensland election, Labor retained the Townsville seat of Mundingburra by just 16 votes. The seat represented the Goss government’s one seat majority, and a series of issues with the conduct of the election saw the Court of Disputed Returns void the result. The Liberal Party won the subsequent re-election and National Party Leader Rob Borbidge replaced Wayne Goss as Premier.
But sometimes the losing party doesn’t challenge. At the 1990 Federal election, Labor lost the Adelaide seat of Hawker by 14 votes but did not challenge. The view was the party couldn’t win a re-election.
At the 1999 Victorian state election, Labor won Geelong by 16 votes. The Liberal Party didn’t challenge, declining support for the soon to depart Kennett government shown by losing three by-elections to Labor in the next year.
If Nicolette Boele’s majority survives a re-count, would the Liberal Party challenge? Whether it could win a re-election in the post-election environment is certain to be a significant consideration.
If Ms Boele does not win the re-count, would she challenge the result? It could be a matter of how much money she and her backers are prepared to spend.
Given how close the result has been in Bradfield, any re-election would be a very expensive affair.
Is the indicative preference count considered accurate, or is it an estimate, such that the full preference distribution would likely change the numbers and potentially the result?
COMMENT: In seats where it is unclear who the final candidates are, the indicative preference count is not very useful. That’s why Calwell remains in doubt as who the final pairing will be is unclear.
The full distribution is usually an opportunity to re-examine and challenge votes. In the end the full distribution is specified in the Act as the official count used to determine the winning candidatre.
Ignoring the subsequent challenging and re-examining of votes, is the indicative preference count expected to be an accurate count of votes as they currently stand, or is it just an estimate of the preference distribution?
COMMENT: It is a count and done accurately but greater scrutiny goes into the formal distribution.
Have you got a vase yet?
Was the best comment made on election night!
Enjoy your retirement.
COMMENT: A beer jug filled in perfectly for a vase.
Is there any time limit on when a liosing candidate can appeal?
Eg Can a party just wait until they improve in the polls – or their opponents decline – and then launch an appeal when they think they have a better chance at winning?
COMMENT: A challenge must be lodged within 40 days of the writ being returned.
Now that the Coalition has split with Liberals holding 17/18 seats, and The Nats on 8 seats: who owns the 16 LNP Qld seats, please? How will they be split? Surely the LNP Qld cannot be a 3rd “opposition” party?
COMMENT: The formation of the LNP split the state’s 30 seats. Of the 16 elected, six represent National seats (Maranoa, Capricornia, Dawson, Flynn, Hinkler and Wide Bay) and 10 Liberals. Some of the sitting Liberals have more connection with the old National Party than the Liberals which is one reason why Queensland has had a conservative influence on the federal Liberals.
TT – Just found this on Bris Times site: “Of the 16 LNP (Qld) MPs elected this month, 10 sit in the Liberal party room in Canberra and six with the Nationals.”
“Former state Labor minister and parliamentary Speaker John Mickel, now a professor of politics at Queensland University of Technology, said, “What [the Nationals] are quickly going to find is, when parliament goes back, they’re going to have to queue up with the independents, etc, to get their question because they’re just another grouping,” he said.
“That will tend to blunt any question time attack because they will have to share some questions with the independents.
“The Libs will be weakened, and they’ll have to share some questions with the Nats. Tactically, that will be a mess after a while.”
If the Libs could manage to be competitive with Labor without the needs for the Nats to officially form government it would be better for them long term. The Libs could then use the Nats on a case by case basis much like Labor uses the Greens, and the rest of the crossbench. By doing so, the Libs could then pursue party policies that are geared to the median voter rather than having to constantly look over their shoulder. Otherwise they will just struggle going forward. The fact is that, while the Nats concerns (and their voters) are valid, the country, and the world, have largely moved on from their positions on most issues. For the Liberals to bend to the Nats will just make them look less serious in the eyes of the swing voters who decide elections. It is largely the reason why the Libs are bleeding seats to Independents. While it is not a perfect alignment, we are going to soon get to a point when the Independent caucus is as large as the National caucus.
Thank you for the explanation.
Is a recount more rigid? It’s the only way my mind would make sense of the rule that it can only be done once 😅I would’ve thought that if a recount resulted in a different outcome than the initial counting, it would warrant another recount.
COMMENT: You mean two out of three to decide it? You have to finish counting and return the writ at some point. And the re-count is done more rigidly because it is the final count and attracts detailed scrutiny throughout. The scrutiny of Bradfield got more intense as the result narrowed. That level of scrutiny will apply throughout in the re-count.
Hi Antony – the best insight into the process I have found – thank you!
“In theory the Court can change the writ to name a new elected candidate, but this is unusual.”
Indeed. There are two instances of this here in the UK that I’m aware of – Mid Ulster in October 1955 and Bristol South East in 1961. The circumstances of each were sui generis, but the common feature was that they both followed by-elections which had returned the same indivduals whose election had been previously voided, despite their still being ineligible. Wikipedia provides further details if other readers are interested.
COMMENT: The Bristol South East got a run in 1996 when Labor argued its candidate should replace Jackie Kelly as MP. It got nowhere as an argument.
There was a third case in Fermanagh & South Tyrone in the May 1955 general election. In all three cases opponents had made efforts to highlight the candidate’s ineligibility to be elected and the media had also covered the cases (two were caused by convictions for an attempted raid on an army base, the other by the sitting MP inheriting a peerage). The election courts were convinced the ineligibilities were clear cut (all had already been through the courts) and that sufficient voters for these candidates were aware of the disqualification and had chosen to “throw their votes away” by voting for them, so awarded the seats to the next eligible candidate.
In all other modern British cases I’m aware of (most election law cases tend to involve local government) the ineligible candidate has been disqualified and a by-election called. Most ineligibilities aren’t even discovered during the election campaign and there aren’t the resources or the media coverage to establish definitive ineligibility then to inform all voters in the time available.
“The Bristol South East got a run in 1996 when Labor argued its candidate should replace Jackie Kelly as MP. It got nowhere as an argument.”
I’m not surprised: the specifics (admirably laid out by Tim Roll-Pickering above) mean it’s just not comparable to other cases and thus can’t establish any kind of precedent at all. (I had misremembered Mid Ulster and Fermanagh & South Tyrone as being one case rather than two)
Though I would’ve thought that the most logical course in the Kelly case, rather than order a new election, would be to recount the original votes with the inelligibe candidate excluded. Presumably, with full preferencing required, every formal vote (including for Kelly) would still have a part to play in the outocome, which is what the system is intended to enable (i.e. no-one’s vote would be ‘lost’ as a result of voting for her). Unless there’s something I haven’t considered?
COMMENT: The argument didn’t stand because in the Bristol South East case, the ruling was the voters elected a candidate they knew was not eligible and the judge ruled the voters had thrown away their votes and appointed the defeated candidate. In the case of Jackie Kelly, the issue of her being ineligible came up after the election, so the voters voted for her in good faith. The debate in Court detailed why the ruling in Bristol South East didn’t apply.
I have very mixed feelings about the citizenship regulations as I can never run for public office because I am a dual citizen. My family are Catholic transplants from Northern Ireland (the Troubles and all that) and I am a citizen of Ireland (and by default the EU). Due to the nature of that situation and sentimentalities attached I am loathe to disavow because I hope to see a united Ireland in my life (I am 44). That said, it has always been something of a curiosity how these eligibility problems seem to surface AFTER the process has begun. Perhaps a lot of them are caught pre-selection and thus not reported but the fact that we seem to hear about them on a semi-regular basis baffles me.
If you are interested in a current case of recount drama look to Canada and the Quebec seat of Terrebonne. After a lengthy judicial recount the victor won by 1 vote, yes 1. As luck would have it, a voter came forth and produced a returned postal envelope for a postal vote that was apparently not counted because of an improper postal code applied to the envelope. To make the story go even deeper, the envelope was prepared by an Elections Canada worker not the voter. Had it been a judge could have easily ruled that it was the voter’s responsibility to ensure the accuracy of their mail. The voter has claimed to have voted for the loser candidate and if her ballot would have been accepted it would have resulted in a tie.
Assuming the two candidates in the Indicative count and the same last two candidates in the full preference distribution count— is there any mathematical difference possible to the result between the indicative and full? Assuming everything was done correctly to begin with…
COMMENT: There are already minor differences. In Bradfield the lead of Boele and Kapterian has constantly bobbed around as ‘a vote here and a vote there’ discrepancies are corrected. In Calwell there have been around five extra informal votes found. These often get found in the distribution of preferences when a ballot paper is examined for next preference.
Apologies if I’ve missed this, but do you have any indication when the recount will begin and how long such recounts usually take to complete?
COMMENT: If as is likely the distribution produces a margin under 100 votes the re-count will get underway as soon as possible.
How is it possible for a candidate to lose votes they have already been credited with ?
Nicolette Boele has just gone back from 190 votes to 186. Gisele remains on 190, as on Friday 23/5/25, c.5.30pm.
COMMENT: It happens when during the distribution of preferences they find either a vote that should have been informal, they find a vote tabulated incorrectly, or they find an error in the original paper work. Scrutineers representing the candidates will be watching all this detail as it is done by the AEC.
I’m curious, is there a check of all the rolls for discrepancies such as duplicate voters? Wonder if a passionate voter might lodge a vote at multiple booths?
COMMENT: All rolls are scanned after the election. Non-voters are sent letters/fine notices. People who show up as having voted more than once are sent please explain letters. Many instances of people marked off twice match up with a similar or roll adjacent name of a non-voter who says they did vote. People marked off more than twice attract greater scrutiny and may be referred to the police. Prosecution is difficult given it is a secret ballot. If there are suspicions of deliberate multiple voting but it can’t be proved to a level where prosecution would succeed, the AEC Commissioner has the power to remove a voter from the printed roll, as is the case with silent voters. Voters not on the roll must complete a declaration vote when voting which ensures they can only vote once.
“If the re-count produces a tie the writ cannot be returned with a named winning candidate and there is an automatic supplementary election.”
If that ever happened then am I right that any new candidates could contest?
If so and the completely new candidate actually wins that electorate then would that be the defining victory for a true believer, or someone who always believed in miracles. Plus a great trivia question.
COMMENT: It is a new election with a new close of rolls and new close of nominations.
“If the re-count produces a tie the writ cannot be returned with a named winning candidate and there is an automatic supplementary election”
its a shame they have something so cumbersome when it comes to resolving a tied result. In the UK they pull a name out of the hat
COMMENT: In an election of 110,000 votes you will find at least one error. Any tie is referred to the Court of Disputed and will go to a by-election unless you can come up with an argument that disputed ballots would have changed the result. But again, some errors can be found in a poll of 110, 000 votes so it would be a certain by-election.
Why are the Special Hospital Team Votes so high for the Liberal Party.
COMMENT: They always are. Nature of the people in nursing homes.
Surprised you didn’t mention the “lost” ballot box in the 2013 WA Senate election.
COMMENT: No ballot box was lost. The issue was the votes were counted but then lost when transferred to archive. Due to a past High Court case, Senate ballot papers have to be stored for 6 years. So after all ATL votes had been counted, they were packed up with all other votes from Forrest, and sent to archive. Somewhere between Bunbury and the archive they were lost.
No one would ever have noticed except that a few weeks later when the Senate result turned out to be ultra close, they decided to do a re-count and all ballot papers were retrieved from storage, at which point is was discovered around 1,200 ballot papers could not be found.
This sort of error does not occur in close seats because all the ballot papers are on the facility where they are counted. The ballot papers counted and check counted for Bradfield remain on premises and are available for the re-count.
Does the AEC attempt to reconcile the total votes (formal and informal) with the number of voters crossed off the roll as having voted? Does it disclose the discrepancy?
I would have thought that, even if no explanation for any discrepancy can be provided by AEC, the mere fact of a discrepancy exceeding the winning margin would be strong grounds for a new election (in that seat).
COMMENT: That is done in every polling place on election night to verify the number of names marked off the roll corresponds to the number of ballot papers issued and the number of ballot papers removed from ballot boxes on opening and counted. This paper work is then check counted the following week as part of the process of finalising the first preference count. Scrutineers observe the process and are informed of any discrepancy. While staff in polling places try to make sure people put their ballot papers in the ballot box, it is not unknown for a voter to throw it in the bin or walk out of the polling place with it. So there can be small discrepencies that scrutineers are made aware of.