Saturday, March 2, 2024

Dunkley By-Election Live

DUNKLEY (ALP, Vic 6.27%)
Jodie Belyea (ALP) vs Nathan Conroy (Lib) and others
By-election caused by death of Peta Murphy (ALP)
CALLED 8:42 pm Labor retain


Updates appear here, scrolling to the top.  When counting is underway refresh every 10-15 mins for new comments.  

Thursday: Labor is now very close to winning the postal count and the 2PP is now 52.71; it is likely to finish between that number and 53.  The Australian published an incorrect article today referring to a 10% drop in turnout; the count does not finish until all postals that can be admitted are received 13 days after polling day.  The turnout is currently 83.5% and there should be about 1% or so to come; the turnout decline will be smaller than at least 16 of the last 20 by-elections, potentially 18.  Media should not publish turnout doomery articles without consulting with the AEC or someone who has a clue.  

Tuesday: With vote totals unlikely to change by even 1% from here it's worth noting an outstanding performance by the uComms seat poll.  I've been critical of poor results from this pollster recently (especially Tasmania 2021) but this one is remarkably good by seat poll standards especially. uComms' numbers with undecided redistributed are below with the actual current numbers in brackets:

ALP 40.1 (41.1)
Lib 39.3 (39.3)
Grn 8.2 (6.3)
LTN 1.6 (2.5)
Ind/Other 10.8 (10.8)

2PP 52 (52.6)


Tuesday: Postals are doing next to nothing now, the 2PP is 52.63 and won't change much from there.

Sunday: There has been a correction in Labor's favour in the Langwarrin booth, taking them up to 52.7.   I have written an article regarding Liberal claims that making this seat "marginal" is meaningful.  History says otherwise.  

10:36 Late night wrap: the loser is ... Advance!  The Carrum Downs PPVC just came in leaving Labor on 52.5, which after perhaps 11,000 remaining postals might make it down to 52 though I would suspect not.  This is a good night for Labor; they've retained a very loseable seat with a swing below the average swing for government vacancies (which is around 6% since Federation).   But for the Coalition it isn't terrible.  The biggest losers here are Advance, who spent heavily on Willie Horton tactics to zero visible effect.  They thereby proved that credit given to their inane behaviour for the Voice result was spurious; the Yes23 campaign defeated itself.  There is a question mark over whether the Liberals have established outer suburbia as a path to victory (if they can't win Dunkley in a by-election, why should they win it in a general?) though the answer may be different outside Victoria.  Labor are right that it is great for them that their primary vote is intact; the final preference flows will be of interest as they do appear underwhelming.   Overall, this is another one of those by-elections like Fadden where one side should be clearly the happier but we can pretty much go back to sleep.  

10:05 Looking at the minors, the Greens result is bad, aside from votes lost to VicSoc and maybe Democrats (though losing votes to a barely living fossil party is inexcusable anyway) I think they've also bled to Animal Justice, Labor and perhaps even Liberals, and partly a result of not announcing a candidate til the last minute perhaps.  Animal Justice have done quite well with a 1% swing.  

9:27 For what it's worth the Carrum Downs PPVC hasn't done anything much either.

8:51 Jane Hume has referred to the fact that no by-election caused by a government death has been lost since 1966.  There have only been five in that time until now.  Three were on margins above 10%, the others being the famous Aston 2001 by-election on John Howard's road back to victory, and a weird one in McPherson 1981 which had had a huge swing to Labor at the previous election.   Sussan Ley has added to the copium with nonsense about first-term by-election average swings - first-term by-elections are usually in opposition seats not government.  Also, a seat does not become marginal because of a close by-election.  

8:48 One thing to see here is that the preference flow to Labor is quite weak.  This is partly the poor result by the Greens but it will be interesting to see the flows.  Their primary vote has held up well.  This is a good result for Labor, the remaining suspense is whether it is OK for the Liberals or bad (my cutoff is 53-47).  

8:42 Called.  

8:40 Dunkley prepoll is in and the swing is not enough there either.  

8:20 The first prepoll has arrived and it is Mt Eliza and there is not enough for the Liberals there. The first lot of postals are also in and the swing there is 5.9% but the swing on early postals is often stronger than on later-received postals.  

8:11 Before this by-election I set 53% to Labor as the threshhold for the Liberal result to be deemed bad.  At the moment we're wobbling around that level but have to see what happens later.  It isn't looking like the Liberal result will be better than OK at any rate.  

7:56 There continues to be some ebb and flow and we will have to wait for prepolls and postals to be sure Labor has won this, though they remain well ahead for now.  Some things are being made of the Greens' indifferent result but they will be dropping some votes to Vic Socialists and Democrats.  The idea that Greens are generally bad in by-elections has not much in it

7:45 Brief dinner break - now things have closed up and there are better booths for Liberals appearing, with the Pollbludger projection down to 52.  

7:20 Now nine booths in and only one of them is strong enough for Liberals so far.

7:16 Another Mt Eliza booth in and the big thing to see here isn't the majors, it's the Greens being clobbered.

7:12 Seaford North, nothing to see here either, a small primary swing that won't amount to much on 2PP.  Overall four out of five booths have been fine for Labor - if things are vaguely close we do need to keep an eye on postals/prepolls to see if they have a different swing.  

7:06 Very variable swings so far - Mt Eliza Central in.  This was the best Liberal booth in 2022 and they've got nothing here, possibly a swing away after preferences. However this booth could be affected by the prepoll location.  I'd be a bit cautious re drawing too much from the Mt Eliza booths yet.

6:59 The 2PP swing in Mt Eliza North came in at 10% to Liberals, so no evidence of weakened flow for Labor in that one.  

6:56 Two more in, a big swing in Frankston Heights East on primaries (but I think not quite enough on estimated 2PP) but nothing much in Carrum Downs.  The idea that Conroy is unpopular in Frankston might be struggling.  

6:47 First booth in, Mt Eliza North. A big swing to the Coalition, 13% on primary and 6% away from Labor, though the right-wing candidates aren't getting a lot.  My 2PP estimate is about the same as the ABC's, about 11%.  Single booths can be unreliable as we saw in Aston.

6:45 Tony Barry on ABC has said that scrutineers say the ALP primary is up but the Liberal primary is up substantially - that doesn't really suggest anything out of the ordinary to me if so.  

6:30 No figures as yet.  

Introduction (5:55 pm)

Welcome to my live and postcount coverage of the Dunkley by-election, a much anticipated test for how the Albanese government is travelling on its own outer-suburban turf.  My preview post is here; it has had little attention in recent weeks as the Tasmanian election hots up but I have just added a few morsels concerning some of the expectation spin from the major parties.  Overall, history suggests that on average chances should be close to evenly balanced here, and a rotten ballot draw for Labor is one dampener on subjective factors in its favour, especially the claimed Victorian Dutton factor.

In the final week the government has had a wobble in national polling and has dropped 0.9% in my aggregate to 51.5-48.5, though 0.4 points out of that is caused by the assumption that Resolve still has a large house effect in Labor's favour.  This week's Resolve was very similar to other polls, so based on past history my aggregate treated it as a great poll for the Coalition.  There could be a bigger wobble if it loses this one.

That said, expectations don't suggest it will lose - the major parties are setting boundaries with Labor claiming that a win by 50% plus one vote is a good result while the Coalition claims holding Labor to 53% will be good (I do not agree with the latter, anything over 53 for Labor would in my view be a recognisably bad result for the blue team).  It has been notable that in the last weeks the campaign seems to have gone off the track of cost of living to a red-meat attempt on immigration by Advance, which probably wants to strengthen its hand as a player by claiming credit if the result goes the Liberals' way.  The Liberals have increasingly bought into this as well.

As noted by Antony Green, the Dunkley prepoll turnout has held up very well at almost the same level as 2022, contrary to earlier by-elections.  Postals are also holding up well with 21983 issued (typically about 14% will not make it back).  The on the day turnout will determine whether the turnout holds up well by by-election standards.  As usual there will be people on social media whinging about the turnout but the final turnout is not known for about two weeks, and turnout should be a complete non-issue in this result.

There are two major prepolls Carrum Downs 12957 votes taken and Frankston 12468 taken (both similar to 2022) and a more interesting moderate sized one Mt Eliza with 4503 taken.  Mt Eliza replaces the Mornington prepoll which was outside the electorate.  It is in a very Liberal voting area and could be a very strongly Liberal prepoll booth - something to keep an eye out for tonight as a Liberal margin of over 1000 votes could occur in this prepoll if the race is at all close.  

A notable change in the field is the absence of UAP and One Nation - this should boost the Coalition primary and possibly also the primaries of Darren Bergwerf (IND) and the Libertarians.  However I expect this to be counterbalanced by a weaker preference flow from the more left-leaning field of minor candidates.  It will be an interesting test of Green preference flows given some of the recent respondent-preference polling.  


  1. Kevin, I am scratching my head and pondering the tea leaves of Dunkley and keep coming back to the question of what swing would be 'average' is key to understanding what happened.

    The former candidate the late Peta Murphy was well liked and has died of cancer leading to this byelection. So first the byelection is a 'no fault' event (not a candidate dropping out because of gun law violations, or bonking, or having a sulk having expected to be in government and now rejects the opposition opportunity). Perhaps the personal vote effect is also null because of the sorry circumstance. And it is a first term of government.

    Is there any meaningful trend based on how many terms, or whether its a neutral or an 'at fault' by election?

    3.8ish is a lot less than 6 if that is the expectation. And a lot more than 1.3% (if that really is the first term of government byelection of government member figures as some outlets are report)

    Thanks for any insight


    1. While one might think that voters would distinguish between bereavement by-elections which are no-fault and resignation by-elections which often do involve fault, historically there is no real average difference between the two. (I haven't specifically tried to distinguish "fault" by-elections because of inadquate historic detail and subjectivity.) An incumbent who passes away still takes their personal vote with them, and if a resigning incumbent was controversial then that incumbent may not have had a personal vote and voters might be pleased to endorse someone else from the party. In terms of term of government, the problem there is that first term government vacancies are very rare events (most first term vacancies are departing opposition MPs of the having a sulk sort that you mention). But it would be expected that governments might do better in their first term to the extent that they are polling better (the fact that the government is polling at about its election level knocks the 6 points down to about low 5s).


The comment system is unreliable. If you cannot submit comments you can email me a comment (via email link in profile) - email must be entitled: Comment for publication, followed by the name of the article you wish to comment on. Comments are accepted in full or not at all. Comments will be published under the name the email is sent from unless an alias is clearly requested and stated. If you submit a comment which is not accepted within a few days you can also email me and I will check if it has been received.