|Yeah OK this is kinda closeish, this one ...|
Cathy O'Toole (ALP) ahead by 8 votes at end (?) of indicative 2PP count
O'Toole provisionally won recount by 35 votes.
Following distribution of preferences O'Toole has won by 37 votes.
O'Toole will be seated and result can be challenged in court, but will retain seat until court (or by-election called by court) determines otherwise.
At least three incidents in which voters were unable to vote have been alleged.
This thread will follow the final stages of counting in the extremely close Queensland seat of Herbert, up to at least the declaration of the result. Herbert was previously covered on the Vanilla Reps seats postcount thread but I have decided to start a new thread for it to prominently address some of the confusion that is going on. However, since I started the thread, more confusion has arisen after the AEC decided to vary from normal procedure in such cases by going straight to a full recount.
On election night it appeared Labor's Cathy O'Toole should win Herbert with a margin of about 50.4:49.6. Then the LNP incumbent Ewen Jones surged in postals to a greater degree than expected based on 2013 patterns, partly because there were more of them. Jones was tracking to win the seat by a few hundred votes with votes running out when two significant counting errors in the ordinary-vote booth counts boosted Labor's position and sent the seat to the wire.
That's where it's stayed, with Cathy O'Toole retaking the lead by eight votes on some of the last remaining votes. The initial 2PP count is over, although at the time of writing 574 votes are still shown on the AEC site as awaiting processing (I understand from the AEC via William Bowe that those are not live votes - they've been rejected and will be dealt with accordingly).
Normal procedure in similar cases
When these two-party-preferred seats are counted, the initial throw of preferences between the top two candidates is for information purposes only. Even if the outcome is obvious, preferences must still then be formally distributed by excluding the lowest candidate and transferring each of their votes to the next-numbered candidate, then excluding the second-lowest, and so on until all the votes are left with two candidates. The end result is a table that looks like this (or did in 2013).
Just about every vote will end up in the same pile (on an LNP vs Labor basis) as the one that it is already in, at the end of this distribution. However, during this process, errors are sometimes found in the original two-candidate count. Also, any count that is this close will by now be intensively scrutineered, with the formality of dubious votes being challenged, and in cases fresh decisions made.
The margin at the end of the distribution of preferences determines whether or not there is an automatic recount. If it is below 100 votes, there must be a full recount; if it is 100 votes or above then the losing party normally needs to show cause why even a partial recount should occur.
The normal procedures are set out in the AEC recount policies. We've just finished second scrutiny (paragraphs 2.3 and 2.4) and third scrutiny (2.5) is completed before the automatic recount can be triggered (4.1).
In the Fairfax count in 2013 (see 2013 late counting thread) there was some movement in votes during this process. Clive Palmer provisionally won by 36 votes at the end of the indicative 2PP throw (the point we're apparently at now). After the distribution of preferences his margin was seven votes, triggering the automatic recount. After the automatic recount, he was declared winner by 53 votes. At that point the LNP could have challenged his election in court but decided not to.
Variance from normal procedure
The AEC has decided to vary from normal procedure and go directly to a full recount, on the assumption that the automatic recount requirement will almost certainly be triggered when the distribution of preferences is complete. This should result in a faster declared result.
What happens after the recount
If after everything we end up with a margin about as close as the current one, then we will have a declared winner. The declared winner will be seated, but may be challenged in court. If the margin is extremely slender the AEC might even request the seat to be voided and a by-election held (multiple voting being one thing that could cause this to occur.) The losing candidate would probably also challenge if there were enough dubious votes for there to be a chance of having the margin, whatever it is, overturned. Depending on the reasons for a challenge, a successful challenge might result in either a by-election or a new winner being declared.
The stakes are rather high, because at stake is the Coalition's ability to have an absolute majority on the floor after providing its own Speaker, and hence suspend standing orders or do anything else that an absolute majority requires, without any crossbench assistance.
But we're not there yet. There is still a long way to go. In 2013 the distribution of preferences in Fairfax took a week and a half, and then the recount (blown out by the vexatious challenging of almost every adverse vote by Clive Palmer's scrutineers) took nearly a month. It will probably all be much faster this time and they should be able to knock it over in time for the August 8 deadline for the return of the writ. The expected schedule is around two weeks from Thursday 21 July.
For those wondering about previous federal close seat records, the closest federal seat margin to stand was seven votes in Werriwa in 1914 but the closest by 2PP percentage was the Griffith by-election in 1939 (8 votes - hat tip to Malcolm Baalman for spotting this one). The closest by 2PP percentage at a regular election was Hawker in 1990 (14 votes). Another famous contender is Stirling in 1974 (12 votes). McEwen in 2007 had a declared margin of 12 votes, but it was amended to 31 by the Court of Disputed Returns, A one-vote margin in Ballaarat (as then spelled) in 1919 was voided by the Court and a by-election held, and likewise for a five-vote margin in Riverina 1903. Both were voided because of electoral irregularities.
Updates on the progress of the recount will be posted here.
Thursday 21 July: Alleged Military Voting Issues: There are claims by Sky News that up to 85 voters based in Townsville and on a military exercise, Exercise Hamel, may have been unable to vote. More details are still needed on this story including how many, if any, were actually enrolled in Herbert. The important point is that should it be found that this is grounds for voiding the final result, then that argument could be made by either party. The fact that soldiers tend to vote conservative is irrelevant. Michael Maley has also noted that given Section 367 of the Electoral Act, a challenge on these grounds might not go smoothly.
Thursday midnight: There is no change in the figures as yet. For the 2013 Fairfax recount detailed special results including numbers of vote challenges by booth were posted but I have seen nothing like that so far in this recount. Hopefully coming!
Friday 11:30 am: There is movement at the station. O'Toole is up from 44,184 to 44,202 and Jones is up from 44,176 to 44,189, meaning that O'Toole now leads by 13 votes. There are 6,444 informals. I've seen reports that what actually happened was that some uncounted declaration pre-polls were added, but didn't scrape enough data to confirm this (see discussion on Poll Bludger thread). I have now copied the current 2PP results by booth so that the cause of further changes can be identified.
There are more claims of irregularities from the LNP camp with Michael McKenna reporting that the LNP claim there were insufficient absent voting papers for Herbert in the seat of Kennedy. I am unsure why this should delay the recount, whatever its implications for the recount's ability to stand.
Saturday 10 pm: I have been offline today but O'Toole's lead is now nine (44195-44186) and there are minor changes in several booths.
Sunday 12:15 pm: Informal continues to make big gains here; Jones currently leads by one vote (44184-44183).
Monday 12:20 pm: Jones now leads by 12 votes (44,191 to 44,179) but there is a great risk of the margin (whatever it ends up being) being overshadowed by other developments. Among these we have the ongoing matter of the ADF personnel unable to vote (because of too few polling booths for their exercise); it is still not clear how many are affected. We have the ballot paper shortage in Kennedy (still no details of the degree of impact there). Irrelevant to the outcome once a court has looked at it, but we also have Labor (and at least one of my commenters!) objecting to the presence of George Brandis as scrutineer. Robert Baird (7 News Townsville) has commented that the AEC expect to finish the recount this week.
Unfortunately we are not getting the kind of rich data on numbers of scrutineering challenges and so on that was posted in the case of Fairfax 2013, which makes commenting on the chances of the margin turning around again more difficult.
Monday 3:15: Jones' lead is now back to 8 votes (44,191 to 44,183). Looking at the booth data, since Friday Jones has made a net gain of 21. 16 of the 42 booths have had changes to one or both candidate totals, with three cancelling out, O'Toole gaining in eight and Jones in five. O'Toole has also gained one absent vote. The main reason Jones has taken the lead is that in a single booth, Railway Estate, he gained 14 votes while O'Toole lost 16. That looks a lot like a small number of votes might have been in the wrong pile there.
Meanwhile there is yet another alleged irregularity with a claim that 39 patients in a hospital ward in Townsville did not get to vote.
Monday 5:00: A massive boost to O'Toole with a 39-vote swing in the booth of Vincent and also a small gain in the Nelly Bay booth, putting her now 37 votes ahead (44209 to 44172).
Monday 5:40: Lead to O'Toole now 44.
Monday 6:50: Things continue moving O'Toole's way - she is now up by 47!
Monday 7:10: Another big gain for Labor, this time a 22 vote margin shift in the Northern Beaches booth. Lead to O'Toole is 73 (44,222-44,149). There are now 25 booths with changes, but a further six are shown as having been updated (presumably without). So there may be only another 11 booths to recount.
Tuesday 11:30 am: Some more slight changes with the lead now 74 (44,221-44,147). No new booths are shown with changes.
Tuesday 11:45 am: Reports that the lead came down to
Tuesday 12:30 pm: The AEC website is back up, eight booths have not had any rechecking now and the margin is 57. (44,214-44,157)
Tuesday 1:15 pm: There are seven booths remaining and the margin is 51. I'm working today so haven't had time to project whether certain kinds of booths are good for one or the other in terms of vote gains.
Tuesday 3:40 pm: Another booth bites the dust and the margin is 49.
Tuesday 4:40 pm: Presumably the smallest remaining booth (a special hospital booth) is in and the margin is 48. Five booths to go!
Tuesday 5:12 pm: The rally by Jones continues but the lead is not coming down fast enough. He trails by 40 with four booths to go. (There may also be postals to be dealt with.)
Tuesday 5:16: Three booths to go and O'Toole has gained a vote and leads by 41.
Tuesday 5:43: Two left and O'Toole leads by 39.
Tuesday 7:21: Just one booth left and O'Toole leads by 37. There have been changes in most of the postcount categories so there may not be much left to go!
Tuesday 7:54: The very last booth has seen a two-vote shift so O'Toole still leads by 35. Awaiting confirmation as to whether that is the end of the recount and at what point a result might be formalised. I would assume after the distribution of preferences.
See the comment by Geoff Lambert who has been tracking the total numbers of votes per booth and has noticed that the changes are not just because of movements back and forth between the candidates and the informal pile, but also the total number of votes at specific booths is changing. I have seen reports to this effect from other electorates too.
Recount Finished: What Does This Mean?
The recount has finished with a margin of 35 to O'Toole. However as confirmed by the AEC on Twitter, this is not the final result. The distribution of preferences will be prepared on Wednesday, commence on Thursday and is expected to take at least two days. Changes are possible in this time but barring extremely sloppy recounting it seems highly unlikely that they will overturn the apparent result.
Distribution of Preferences
Thursday 12:00 The distribution of preferences is presumably underway since O'Toole has gained a vote and now leads by 36.
Friday 11:00 Only two booths have been changed so it seems very few errors indeed are being found and O'Toole now leads by 35.
Saturday 1:30 Eight booths now have changes and O'Toole leads by 39. Reports vary as to when the distribution will finish with some saying today and others saying the result will not be known til early next week but it looks very much like only small errors are being found, of a vote here and there, and that O'Toole is going to be seated.
Sunday 1:00 The distribution of preferences has finished and Cathy O'Toole has won by 37 votes. She will be declared the winner and is the new MP for Herbert. It remains to be seen if a legal challenge will be lodged in the 40 days following the election, but in the event that there is a challenge, O'Toole will be the MP for Herbert until such time as the court either:
* reverses the outcome and awards the seat back to Jones, or
* orders a by-election which O'Toole then fails to win
The first would only happen if the court reversed formality rulings about a large number of votes and as a result declared Jones the winner, but this is not likely. In the 2007 McEwen case, court rulings on vote formality affected the margin by 19, but awareness of those rulings should mean that such a large impact is unlikely.
The ball is now in the Coalition's court as to whether they bother to challenge.