Saturday, March 28, 2020

Queensland: Bundamba and Currumbin By-Election Counts


BUNDAMBA: McCallum (ALP) retain
CURRUMBIN: Gerber (LNP) retain

Friday: With both seats decided I am paying little attention to the counts but note that the error mentioned below is now corrected.

Tuesday night: As Antony has noted, the preference figures that have gone up for Bundamba appear to have transposed the flow of Greens preferences making them flow 2:1 to One Nation instead of a more logical 2:1 to Labor, so McCallum will get more like 59% than the 54.9% he's currently credited with.

Tuesday 4:50:  Antony Green has reported that the LNP are around 500 votes ahead with not enough left to overturn.  These numbers are still not showing on the ECQ website.  Labor has conceded and the LNP has claimed Currumbin.  There will be a swing against it which may be close to 2% but is to be determined.

Tuesday 1:40: There's nothing happening here, what it is aint exactly clear ... no change on yesterday's figures for Currumbin.

Monday 11:50: The "unofficial indicative count" of 2PP (hmm, I thought they were going to start the official count yesterday) has commenced and is currently showing Gerber ahead with 51.2%, however that is off only 17.9% counted.  We don't know where these votes are from, though it may be they're nearly all the booth votes.  As more votes are added this indicative count may be bouncy depending on the order in which parcels are added, so if Campradt moves ahead in this count at some stage nothing should be read into that until the numbers settle down.

Monday 9:30: William Bowe now has a full booth breakdown which shows that all that is left to count here are out-of-electorate prepolls and absents, remaining postals, and presumably a few provisionals.  Interestingly the pattern here shows little difference between on-the-day voting and prepolls,  Here is a table with 2PP estimates based on the flows posted by Antony Green:

(Other is probably mostly telephone voting.) There doesn't seem to be serious doubt here that Gerber has won.  ABC estimates 51.2% and late postals should assist Gerber (though not as strongly as the early postals), but I'm still waiting til we see some real 2PP numbers to be sure.  There might not be any swing though I think there will be a small or perhaps very small swing to Labor.  The LNP would be nervous about the seat at the main election based on this.

Sunday overnight: Antony Green has reported weak scrutineering preference flows to Labor from the Greens (only 71%), further suggesting that Gerber should win Currumbin, albeit probably with a small swing to Labor.  Meanwhile the Premier has criticised the ECQ's failures with result reporting at this election, but it would be interesting to know how well the ECQ is funded by government and whether that had much to do with it.  The ECQ's performance at the 2017 election was also less than ideal.

Sunday 9:10 William Bowe has now posted booth figures for most of the votes counted so far.

Sunday 5:00 A very slight update with a further fraction of a percent added and Gerber's lead up by 0.04%.  XML still badly incomplete.

Sunday 3:10 A sudden jump to 68% counted.  Gerber 43.87, Campradt 38.62, Spain 10.63, Bettany 6.88.  So Gerber is up 1.6 points on before while Campradt is down 0.7 and Spain is down 1.1.  On this basis if the ABC's preference flows are accurate then Gerber is probably slightly in front already.  The XML file includes votes for various on-day booths and prepolls but no postals, but it is missing about 9000 of the nearly 22000 ballots already counted, and I expect these would include some postal votes to hand so far.  Votes still to come would at least include postals not received yet, and I would expect them to also include out-of-division prepolls and absents (assuming absents were taken).

Of note is a high informal rate in both counts - Currumbin 7.6%, Bundamba 10.8%.  In 2017 informal rates for these divisions were 4.65% and 8.24%.  Probably these gaps will narow as more postals are counted (they tend to have low informal rates), and perhaps on ballot checking.  In the event the informal rates remaining significantly above 2017 there are two possible issues here: firstly a protest vote against the elections being held at all, secondly confusion between the compulsory-preferences by-election and the optional-preferences council elections on the same day.  Detailed study of the ballot papers might be needed to determine which (or perhaps looking at informality changes for the council votes would shed some light on the protest vote possibility.)

Sunday 2:05 Still nothing.  I did set a low bar for the progress of this count at "better than Iowa" but that may have been unduly optimistic.

Sunday 11:30 Nothing has happened on the ECQ website in the last three hours except that the Bundamba count has advanced slightly.  I am keeping an eye out for any useful information through other channels but haven't seen any.

Sunday 8:30 More figures have been posted overnight.  In Bundamba with 56% counted McCallum continues to lead on 42.88% from Bell (One Nation) 27.8%, Shearman (LNP) 15.92% and Mutton (Green) 13.4%.  Bearing in mind that the final count will probably only reach 80% counted, not too much will change from here - the parties will finish in that order and McCallum will have no trouble winning.

In Currumbin, with 37.11% counted Gerber (LNP) has 42.28%, Campradt (Labor) has 39.36%, Spain (Green) has 11.73%, Bettany (One Nation) has 6.62%.  Campradt would very likely win if it stayed like that - the ABC estimates 51.1% 2PP - but the problem is not knowing how representative those numbers might be.  The current figures must be a mix of on the day and pre-day votes, but there is no breakdown available.  Presumably on-the-day votes are overrepresented. If postals were also overrepresented, the count could be fairly representative overall, but if the current count is mostly a mix of on-the-day votes and prepolls, and all or nearly all the on-the-day votes are in, then I would expect Gerber to win.  Also given the lack of details it is hard to be sure there are no further errors as were seen early in the count.  This one will take a while ...

Saturday 10:10 I'm calling stumps on this one for Saturday night and will resume tomorrow once there are more figures or other significant news up.

9:39 The ECQ has acknowledged that its computer system has failed: "We're having  technical issues displaying results online. We are are working on the issue."  Not sure we will see more tonight with the unofficial count expected to end around 10, it might be a case of come back tomorrow.

9:24 Antony again: "From a hand-scribbled A4 sheet, it seems the LNP leads Currumbin 3200-3167 from 10 counting centres, but 12,000 pre-polls to be counted and then the LNP leaning postals after that. That looks like a swing to Labor, but unreliable given how few votes were cast on the day." I'm unsure if he means 2PP or primaries; even if these are primaries the LNP would still be competitive once its likely edge on prepolls and postals was factored in.

9:20 There is another website issue with the Brisbane City Council mayoral count where it seems that votes for the Greens have now been miscredited to the candidate below them, curiously changing their percentage of the "vote" but not his,

9:10 Antony Green: "I hear there are lots of results being counted and phoned through, but when it comes to letting the world know them, it's a case of "Computer says no". Technical problems."

9:00 Still nothing.  A fair amount of progress in the local government counts though.

8:30 Nothing to see in either seat for quite a while now.  It is unfortunate that we have an online count crashed on a night when no scrutineers are allowed to watch the booth counts, since otherwise I would expect scrutineering reports to filter through and fill the vacuum.

8:16 The only good news for the LNP in Bundamba is they seem to have avoided coming last.

7:51 Still waiting for Currumbin to be restored to consciousness - it's been taken out of the XML file.  The swing in the Currumbin booth, if correct, was about 13 points off LNP and 10 to Labor.  Note that the ABC website's preferences for Currumbin are estimates, not real numbers (and we don't know if the primaries are real numbers either.)

7:35 Bundamba is clearly Labor vs One Nation. We now have the Bundamba Early centre in and in this McCallum (Labor) leads 42.0-30.3.  That is a very strong sign for McCallum.

7:30 There appear to be nine booths in in Bundamba and I can only match five of them based on booth names but those I can match seem to be representative.  So McCallum seems to have a strong lead (now 41.9-28.2) in on-the-day votes.  With the issues I mention below in Currumbin I would read nothing into the numbers there as there may be serious issues with them.

7:21 In the XML file there's an issue with the Palm Beach South booth in Currumbin which is showing votes only for Gerber (LNP).  In the Coolangatta booth she leads 48.4-34.1 (very little swing) while in the Currumbin booth it's 32.7-50.3, a very large swing to Labor in that booth if accurate.

7:16 Now Currumbin with 3.5% counted has the LNP candidate on close to 50%, which is a good start unless they come from somewhere very unrepresentative.

7:15: All coming in a rush, the XML shows that a large number of booths are now included in the total so what we are seeing now is probably a pretty fair idea of where the on-the-day vote finishes up.

7:10 The XML feed is yet to show where these Bundamba numbers (Labor now 41.6%, ON 26.2%, Green 16.3, LNP 15.8%) are from.

7:06 First figures from Bundamba, with Labor well ahead, just finding out where these are from.

6:50 First figures expected soon, "from around 7pm".

6:40 While we're waiting, a note about the ECQ's "unofficial" vs "official" count lingo.  This has resulted in some comments to the effect that the unofficial results are not worth covering, but this is just a normal election-night check count process, albeit one that might be more error-prone than usual without scrutineers there to catch out mistakes.

6:05 (Qld time) Results pages are up here: Bundamba, Currumbin.  Nothing to see yet however.


I've put up a page here for count updates on the Bundamba and Currumbin state by-elections in Queensland, both of which (controversially) had their nominal polling day today.  Bundamba (21.6%) is a nominally ultra-safe Labor seat where One Nation might be a threat, while Currumbin (3.3%) is an LNP marginal.  Both have been vacated by long-term and somewhat maverick incumbents who are estranged from their respective parties.  I don't intend covering the Queensland council elections as well - for which see Tally Room , Poll Bludger and Antony Green - but I may add the odd comment about them.  I earlier posted some background here.  If anything, events since that piece was written have probably lowered the stakes for the LNP in Currumbin, since a loss to the government could now be explained away as a rally round the flag in a time of crisis (whether that was actually the case or not).  Sportsbet has the LNP at 1.50 to retain Currumbin to Labor's 2.50.

These by-elections are of interest especially because of the impact of the current coronavirus on everything to do with them - political messaging, when and how people vote, how the count proceeds, and how bad the turnout is in a nominally compulsory election.  Huge numbers of voters have already taken advantage of postal voting and early voting.  It is to be expected that elderly voters in particular will not be voting much on the day - as a result on-the-day votes from booths could be unusually unrepresentative.  Estimates I've seen suggest about 18-25% will vote on the day, with most voters prepolling.

I don't know how slowly the count will proceed but I'd expect the booth counting tonight (which I understand will go ahead, albeit without scrutineers) to be at least slightly slower than normal. However the ECQ does say that preliminary counting of on-the-day booth votes, prepolls and postals will all occur from tonight and will be posted.  With large prepoll booths we will just have to see how that goes.  There will also be preference counts.  Post-counting is also likely to be slow based on an ECQ directive re scrutineering.  I'm not expecting these counts to be an Iowa-level debacle, but don't expect too much to happen too quickly.

There have been anecdotal reports of voters not receiving requested postal votes and speculation that this could result in a challenge to any micro-close or even vanilla-close result.  I'd actually be surprised if anyone bringing a challenge managed to get it decided before parliament rises for the election in October, but we'll see.

I'll be posting tonight as results come in but will be slow on anything that happens in the first half hour because I'll be getting dinner.  I'll then follow the counts in coming days to the extent that the results may still be in doubt or otherwise of interest.  Assessments are all provisional until a seat is CALLED, an action that may be taken with even more than my usual level of reluctance.

No comments:

Post a Comment