Category: Queensland By-Elections

1.00am. The ECQ uniquely publishes preference flow data by candidate on the night, from which we learn that 86.4% of Greens votes favoured Labor over the LNP despite the lack of an active recommendation on their how-to-vote card. Labor got 81.6% of preferences from Liam Parry of the unregistered Queensland Socialists; 60.2% from Legalise Cannabis; 24.4% from Family First; 62.7% from Animal Justice; 36.5% from independent Damian Smart; and 33.6% from Libertarians.

10.56pm. I’ve finally fixed what was screwing up my probabilities, which were essentially behaving as if only a few hundred votes had been counted. So it is no longer countenancing the possibility of a Greens win, which was obviously never on the cards, and getting towards calling it for Labor without going all the way, as I believe is appropriate.

10.40pm. With little further ado, its TCP results are now in, and they added only 181 to the Labor lead, though at 768 votes it’s almost certainly enough.

10.30pm. At long last, the Chermside South pre-poll has reported its 6833 primary votes, which have both major parties higher and the Greens lower. If the preferences flow as they have been elsewhere, they should add about 300 votes to a Labor TCP lead that’s currently at 587, which should settle the deal for them.

9.21pm. I don’t have all the data on this I would like, but I would roughly guess that there will be about 3000 late-arriving postal and the LNP will make up about 400 to 500 votes on them. The only other category of vote outstanding after Chermside South pre-poll is in will be in-person declaration votes, which favoured Labor 267-138 in 2024.

9.21pm. The TCP count has caught up with the primary vote count, and Labor has a seemingly handy 587-vote lead. However, neither result has reported from the Chermside South pre-poll, which Labor won 54-46 in 2024 compared with an overall result of 55.3-46.7. If Labor can hold their ground there, they can probably be favoured to hold out against the late-arriving postals.

9.11pm. The Stafford pre-poll has reported its TCP result and it’s pushed Labor to a 121-vote lead, though they have in fact gone backwards slightly on my projection, which has them winning 50.4-49.6.

9.07pm. The LNP has a 17-vote lead on the TCP count, with three election day booths plus the Stafford pre-poll only having reported the primary vote. Those booths are unlikely to change the TCP much when they report. The big question is whether Labor does well enough on the Chermside pre-poll when it reports to give them a buffer sufficient to hold out against late-arriving postals, which will assuredly favour the LNP.

8.25pm. I should also note that I rolled the Brisbane CBD pre-poll booth, which is not in operation at this by-election, into the Stafford pre-poll result for purposes of historical comparison, which was a) a bit inelegant, and b) gave that booth a baseline for comparison that was stronger for Labor than it should have been, since they did well on the Brisbane CBD result. This means the swings from the booth are a bit better for Labor than the results just cited, and gives them hope for a strong pre-poll dynamic that will carry over to the Chermside pre-poll when it reports. On the other hand, Labor won absents 57-43 in 2024 compared with 55-45 overall and there will be none of those this time, an anomaly my system doesn’t factor in. So swings and roundabouts, in other words.

8.18pm. The Stafford pre-poll booth, with 6254 votes, has gone as Labor might have hoped, with Labor down 7.5% compared with 10.8% on election day booths, and the LNP up 1.9% compared with 3.6%. Which at the very least keeps them in the race.

8.05pm. The postals are now in on TCP, and they have indeed pushed the LNP to a quite handy 52.7-47.3 lead that the remaining election day booths are unlikely to change much. My system is likely flattering Labor a little because they did well at the 2024 election on absents and the Brisbane CBD booth, which aren’t in play this time. They can still win if they do well on pre-polls, which may not report until later in the evening, but you would rather be the LNP at this stage.

7.52pm. 3656 postals have just unloaded, and they have swung more gently than most of the rest of the votes so far (Labor down 5.6%, LNP up 4.8%), such that my system now has it neck and neck. They are still strong for the LNP in absolute terms though, as postals usually are, so they may have the effect of pushing the LNP into the lead on the TCP count (on which they continue to trail) when they report their TCP result.

7.39pm. Twelve of 14 election day booths are now in, and all you can say with confidence is that everything is riding on the pre-poll booths, which may not report for some time.

7.38pm. TCP blockage cleared, and now my system is projecting preference flows off an actual TCP count, Labor is getting 70% of them rather than 67%. That’s far from decisive, but it slightly moves the needle in their favour.

7.34pm. The Grange booth was one of Labor’s less bad results. Touch wood, the TCP results blockage should be fixed in the next update.

7.30pm. Actually, it looks like there are TCP results but my system is failing to read them. Looking into it.

7.27pm. The Stafford Heights booth is slightly better for Labor than its performance generally so far. The LNP is still slightly favoured, but the uncertainties noted in my 7:17pm update remain. Still nothing on TCP — my estimates are giving Labor 80% of Greens preferences, which is less than I’d normally have because of their open HTV card.

7.24pm. Chermside East is a bit better for Labor than the other two Chermside booths, but everything in the previous update’s assessment still stands. Still nothing on the TCP count.

7.17pm. Chermside West booth almost as bad for Labor as nearby Chermside South. Labor’s hopes are for a better flow of preferences than I’m estimating, a stronger dynamic in booths nearer the city that are yet to report, and a better trend on pre-polls.

7.12pm. The large Kedron West booth is less bad for Labor than Chermside South, but still bad enough — probably enough for them to lose narrowly if it proved representative, and certainly not good enough to get them back ahead on the projection. A lot depends on preference flows though, and there is still nothing on TCP.

7.06pm. Bad result for Labor from Chermside South booth gives the LNP a projected lead. A lot depends here though on the accuracy of my preference estimates, there still being no TCP results.

7.00pm. The first actual booth to report is Stafford West, and it’s living up to suggestions of a tight result: Labor down a dangerous 8.5%, but LNP only up 1.8%. Still nothing on the TCP count.

6.49pm. Three results in on the primary, none of them from static booths — telephone voting, pre-poll telephone voting and mobile voting. Consistent with a close result, though I’m projecting a narrow Labor lead. This is based on preferences estimates though, in the absence of any numbers on the TCP count.

6.37pm. Another minor result, 308 votes from Telephone Voting, is better for Labor, the LNP swing being only about 4%. My system has been a bit slow to update so far due to bugs, but I should get them sorted shortly.

6.30pm. The Mobile Polling booth result came through promptly, with only 106 formal votes, but they’re encouraging for the LNP as far as they go with a 12.7% increase on their primary vote and little change on Labor’s.

6pm. Polls have closed for Queensland’s Stafford state by-election, first results of which are likely to come through in about three-quarters of an hour. For what it’s worth, a Courier-Mail straw poll of 150 voters conducted at booths this morning found 43.3% saying they had voted LNP and 32.7% Labor, which would point to an LNP win if accurate, though the margin of error would have been 8% even allowing for a genuinely random sample.

The third by-election in as many weeks takes place in Queensland on Saturday for the state seat of Stafford, vacated by the death on April 4 of Labor-turned-independent member Jimmy Sullivan. The by-election is a return to simpler times in being a straightforward contest between Labor and the Liberal National Party with the Greens as the biggest party on the sidelines, One Nation having opted not to run in a seat that is too close to the city to be convivial territory for them.

The question is thus whether a Labor opposition enduring the doldrums that typically follow being evicted from office can defend a margin of 5.3%. David Crisafulli’s government is at a correspondingly favourable point in its life cycle, although it has been enduring an ill-timed ministerial scandal over recent weeks. The Greens have placed a speed bump in Labor’s path by declining to direct preferences on their how-to-vote card — this typically makes a difference of about 5% to the flow of their preferences to Labor, which is non-trivial in the context of a seat where its vote share at the 2024 election was 18.1%.

The closest we have come to an opinion poll is a straw poll of 300 voters conducted at the electorate’s two pre-poll centres by the Sunday Mail, which found 41.7% saying they had voted for the LNP, 36.7% for Labor, 12.7% for the Greens and 8% for independents. This compares with 2024 election results of LNP 38.1%, Labor 38.8% and Greens 18.1%, or 39.6%, 39.3% and 17.0% specifically among pre-poll votes. The accompanying report oddly claimed this put the LNP “on track for by-election upset”, despite the movements being inside the Labor margin. The independents include Liam Parry, part of the unregistered Queensland Socialists, who has been in the news as the first person charged for using the proscribed “from the river to the sea” chant at a pro-Palestine rally in March.

Nine Newspapers’ online publication the Brisbane Times has the bi-monthly Queensland state poll from Resolve Strategic, combining the Queensland responses from its last two monthly polls. It records a four-point drop in support for the Liberal National Party government, now on 30%, which has unusually little to do with One Nation, who are up only one point to 17%. Labor is up two to 28%, and the Greens are up one to 11%. David Crisafulli retains a strong net likeability (meaning favourable minus unfavourable, and ignoring neutral) of plus 19, down from plus 21, while Steven Miles is down from minus three to minus five. Crisafulli leads 42-26 on preferred premier, in from 44-23. The poll was conducted between March 8 to 14 and April 12 to 18 from a sample of 870.

We will perhaps get an insight into the accuracy of all this when the Stafford by-election is held on May 16, which has been brought on quickly following the death of Labor-turned-independent member Jimmy Sullivan on April 9 – although the contest has been deprived of one point of interest by One Nation’s announcement that it will not field a candidate. The two major parties have moved quickly to endorse candidates: Labor’s is Luke Richmond, the party’s assistant state secretary and a former ministerial chief-of-staff and official with the Right faction Australian Workers Union, while the LNP has again endorsed its candidate from 2024, Fiona Hammond, who formerly served Marchant ward on Brisbane City Council and has more recently been government relations manager at an engineering consultancy. The closure of nominations and ballot paper draw are on Friday.

End of evening. With only late-arriving postals and provisional votes to be counted, Liberal National candidate Wayde Chiesa holds an insurmountable 3.8% vote ahead of Katter’s Australian Party candidate Mark Molachino, whose party’s vote is down from 46.4% to 30.2%. As was noted often during the commentary below, large regional effects were evident: the two-party swing was 12.4% in the booths at the Townsville end and 30.0% further north. The KAP two-party vote was 11.1% stronger at the former end this time, whereas at the 2024 election it was 6.4% weaker. Support for One Nation remained fairly modest despite almost tripling on the 2024 result, while the Labor vote was well down off a low base, making their determination to field a candidate all the more puzzling.

8.48pm. Primary votes are in from the two pre-poll booths, and the one in Ingram (“Returning Officer Hinchinbrook”) has delivered what my system deems to have been a killer blow to the KAP: out of a substantial 5348 formal votes, the KAP is down there 38.8% and the LNP up 31.2%. The Deeragun early voting centre, at the Townsville end, records a swing of only about 10%.

8.32pm. I now see the real reason my projection ticked back to the KAP, namely that 3393 postal votes reported. However, it’s now tipped back over 80% LNP win probability, I think because of new TCP numbers resulting in a revision to the preference projection. Still nothing from the two big pre-poll centres.

8.25pm. The latest update caused my LNP win probability to fall quite substantially to 74.4%, which I think is down to one TCP result having given the KAP a better flow of preferences than elsewhere. After a fair bit of variability, the ABC’s and my TCP projections are now in accord at 52.1-47.9 in favour of the LNP.

8.13pm. TCP results from seven booths have dropped, without changing the outlook much.

8.03pm. Regarding those preferences: it may be noted that One Nation’s how-to-vote cards favoured the LNP over the KAP.

8.00pm. Two substantial booths have reported on the TCP, greatly improving the data from which to project the outcome, and my system is interpreting it as good news for the LNP, whose win probability is back well above 80%. I note that the ABC’s projection, which actually had the LNP falling slightly behind for a moment there, has now substantially revised in their favour — presumably because it’s no longer going off preference estimates that may have been flattering the KAP. The note of caution surrounding what the two big pre-poll booths might do remains relevant though.

7.46pm. With most election day primary votes in now, the race has tightened considerably — early projections based off big swings in rural booths have proved deceptive. Still more data needed on TCP, though the biggest imponderable is probably the two big pre-poll centres.

7.36pm. There are now five booths in from the Townsville end, and while they are less bad for Katter’s than booths further north, they don’t quite get them where they need to be. Their remaining hopes involve a stronger result on pre-polls and a better flow of preferences than is currently being projected off a low base of data (only “Telephone Voting – Early Voting” is in on TCP).

7.25pm. My projection is now going off actual preferences rather than estimates — specifically, by projecting from one booth where 149 votes slightly favoured the LNP. Which suggests my preference estimates weren’t so far off the mark, and pushes the LNP win probability back over 90%.

7.18pm. I’ve just revised by Labor preference estimate in favour of Katter’s, so expect improvement in their favour with the next results update.

7.18pm. Ten booths in — I’m projecting an LNP TCP of 58.1% but the ABC’s only gets it to 54.4%. The difference is our respective guesses about how (mostly) Labor and Greens preferences will flow.

7.16pm. Nine booths now in on the primary vote, still nothing on TCP, and the swing has moderated a little – my projection of the LNP primary vote is in from a bit over 47% to 43.6%. It may be notable that the difference has been made by the first booth from the Townsville end of the electorate (Deeragun), where the swing is about 10% and not as high as 30% as it’s been in some of the rural booths. The possibility that the Townsville booths will maintain that pattern means it can’t be called yet. One Nation look too far behind the KAP in third to be a shot of making the final count.

6.54pm. After what seemed to me a slow start, we now suddenly have six small booths in on the primary vote, with hugely encouraging results for the Liberal National Party – there were strong Katter areas, but their vote has fallen by more than half, with a lot going to One Nation and slightly more to the LNP. One Nation have at least a prayer of getting ahead of the KAP, and I’m not ruling either of them out yet, but unless we’re seeing substantial regional variability this is looking like an LNP gain.

6pm. Polls have closed for the Hinchinbrook by-election, occasioned by Katter’s Australian Party member Nick Dametto’s departure for a successful run at the Townsville mayoralty. The electorate contains a number of small rural booths, so first results should presumably in well within the hour.

Tomorrow is polling day for Queensland’s Hinchinbrook by-election, which will fill the vacancy caused by Katter’s Australian Party member Nick Dametto’s successful bid for the Townsville mayoralty. Dametto’s move was launched with evident confidence that the party could retain the seat in his absence, but their candidate Mark Molachino has faced a determined campaign from the Liberal National Party. The by-election also offers some sort of litmus test for the resurgence of One Nation, which made little headway against Dametto at the 2024 election, polling only 4.6% of the vote. Perhaps surprisingly, Labor is also making the effort to field a candidate in a seat where they scored 14.0% of the vote last year.

The LNP campaign has made much of Molachino’s past as an ALP member, and also of Premier David Crisafulli’s connections to the electorate, which encompasses his old home town of Ingham. Crisafulli is notably close to his party’s candidate, Wayde Chiesa, the connection extending to his role as chief financial officer of Southern Edge Training, of which David Crisafulli was the director until two months before it collapsed in 2016. Both have been muscling up on law-and-order, the LNP campaign emphasising tougher bail laws and the KAP promoting its showpiece “castle law” policy, which would allow residents facing home invasions to use “whatever force necessary”.

As reported in The Australian’s Feeding the Chooks column, One Nation’s how-to-vote card is notable for having Chiesa ahead of Molachino, which high-profile One Nation figure James Ashby attributes to “former Labor Party people always doing the dirty on us”. The LNP has One Nation ahead of the KAP, while the KAP card is split between options favouring the LNP and Labor, both with One Nation second. Also in the field are Amanda Nickson, whose how-to-vote card order boils down to Katter’s, One Nation, Labor, LNP; independent Steven Clare, who was One Nation’s candidate for Thuringowa at the state election; and Aiden Creagh of the Greens.

As always, this site will feature live reporting of results and commentary from the close of polls at 6pm tomorrow (7pm AEDT).