Sunday, May 22, 2022

2022 House of Reps Postcount: Macnamara

Macnamara (Vic, ALP vs Lib 6.3)
(2019 3CP Lib 39.5 ALP 33.2 Green 27.3)

Exclusion order issue: final two undecided with Liberal, Labor and Greens all very close at 3-candidate point.

Labor needs to remain in top three to win

Labor will win the seat - just!

This post will follow the postcount in Macnamara, which under its old name of Melbourne Ports delivered the most exciting postcount of the 2016 election, with a two-way contest between Labor and the Greens to make the final two and the Liberals also a chance to unseat Michael Danby.  Six years later the Liberals have totally cooked their chances in the seat after picking two dud candidates in a row, and Josh Burns has replaced the more controversial Danby and appears popular and to be set for life.  Or is he?

As it turns out, not so simple, because the excellent result of the Greens generally in this election has seen them pick up a large primary vote swing at the Liberals' expense, while Labor has only a small one.  With ordinary votes plus the first 1998 postals in, the standings are:

Burns (Labor) 32.14%
Hodgins-May (Green) 31.45%
Harkin (Lib) 27.13%
Lib Dems 2.42
UAP 2.02
Myers (IND) 1.73
Animal Justice 1.73
One Nation 1.38

In 2019 the following pertained to the postal, out of division and absent votes for Macnamara:

* On postals the Liberals outperformed their ordinary vote by 15.2%, Greens underperformed by 10.5, Labor underperformed by 6.4
* On declaration prepolls the Liberals outperformed by 1.7, Greens underperformed by 0.7, Labor underperformed by 2.8
* On absents the Liberals underperformed by 1.5, Greens overperformed by 2.9, Labor underperformed by 3.3

The postals effect is a result of the division's large conservative Jewish vote, which often makes it the division (alongside Flynn) with the largest postcount shift to Coalition on a 2PP basis.  In 2019 this shift was 2.7% 2PP.  We can see this in the early postals counted where Liberals are outperforming by 13.6, Labor is outperforming by 2.2 and the Greens are underperforming by 16.1.

In 2019 about 6% of Macnamara voters voted by out of division prepoll and about 5.5% absent voted.  This year the rate of postal votes mailed was 22%, but of these only around 75-80% will be included.

On a rough model from last year's patterns and ignoring patterns in the postals received so far, I project that you could throw a blanket over the three at the end of the primary count (my very rough estimate is Labor 30.3 Greens 30.2 Lib 29.7).  If that was so, we would need to see if minor party preferences were capable of putting Labor into third.  If Burns is in the top two, he wins, as the Liberals recommended Labor over the Greens on their card, which tends to produce about a two to one flow.  

However, the micro-party mix includes Liberal Democrats, UAP and One Nation,  In 2019 the Liberals gained on Labor and the Greens by about 35% of the UAP's vote share, and the Greens gained on Labor by about 35% of Animal Justice's.  On the projection there is a serious chance that Labor would be pushed into third.

That said, the early postals are more favourable to Burns than the projection, in all respects (worse for the Liberals and Greens, better for him).  This may be just a feature of early postals, but it may carry through to later counting (especially as Labor is not normally strong in early postals.)  If Burns can continue to outperform the postals projection from 2019 then he should easily beat one of the other two and be returned.  At the moment I think this is likely (he can beat either of them) but there is enough in this to consider the seat not yet bolted down.  I will look at this in more detail over coming days but I think that is enough to start the ball rolling on this one on.  

Monday night: More postals are added and now the Liberals are outperforming by 14.5, Labor by 0.6 and the Greens are underperforming by 15.1.  Current standings are Burns 32.16 Hodgins-May 29.92 Harkin 28.66.  There are still something like 7000 formal postals to come.  

Assuming the current postal rates continue then I am projecting Burns to about 32, Harkin 29.8 and Hodgins-May 29 after absents and dec prepolls.   At a very rough estimate of the minor preference flows I then project Burns and Harkin on about 34 and Hodgins-May on 32.  Burns only has to match my projection or come near to it because if he gets over a third on 2CP he wins.  Even if he underperforms he can only lose if Hodgins-May also outperforms Harkin on the projection, Harkin outperforms Burns.  So it seems Burns is still fairly well placed but sheesh, this is complex!

Wednesday: There are some preference estimates posted by Antony Green with the Liberals getting 48% of preferences, Greens 34 and Labor the rest.  Applying this to the current live count I get both majors around 33.8 and the Greens around 32.4.  Projecting to the end of the primary count and assuming remaining postals will be somewhat weaker than those so far I get Liberal 34.2 Labor 33.5 Green 32.4.  Even assuming the remaining postals do nothing I only get the Greens up to 32.8.  So on projection I have the Greens several hundred to over 1000 behind after preferences, which would require an unusually strong relative performance on dec prepolls and absents to overturn - or counting errors.  The advantage remains with Labor but it's not yet quite in the bag.

Thursday: A lot of postals have been thrown with now just over 1000 awaiting processing.  As usual the flows on late postals continued to weaken.  After preferences my projection is now Liberal 34.1 Labor 33.4 Green 32.4 so not substantially closer than yesterday (I am assuming about 500 postals still to arrive; there might be more but by this stage they'll be splitting pretty evenly anyway).  This throw has also removed the possibility that the postals uncounted as of yesterday would be much better for the Greens.  For the purposes of monitoring future postals, so far it's Liberal 6890 Labor 5572 Green 3408 from 17559 formal postals.

Friday: The AEC has released a partial 3CP of Macnamara with the Greens only 407 behind (Labor 15408 Green 15001 Liberal 14895) but this is based off a 48% count and I am not sure of the full list of booths/parcels included.  If it's a representative sample that would project to the Greens being, say, 750 behind which could get very close on absents).  More information if I can get it but for now caution is advised.

Friday midnight: 1000 dec prepolls were counted today and they were very strong for the Greens (32.3%) and weak for the Liberals (23.0%).  Ignoring the AEC data which is not necessarily usable, my own back of envelope also projects a c. 700 vote margin to Labor.  This is assuming that the flow on dec prepolls is somewhere between what is seen so far and expected from last time (because we don't know where these dec prepolls are from) and that the flow on postals slows down a lot too.  However, I'm not including COVID phone votes of which we know nothing.  

Saturday: First absents break slightly more strongly to Greens and about 6 points weaker to Liberals than projected. When home a bit later I will estimate a notional current margin as distinct from the projection above.

Saturday 4 pm: Antony Green has released a 3CP with Labor 26782 Green 26528 Liberal 26213, which is missing about 3000 postals.  We don't know yet if these are typical postals of the strong for Liberals, weak for Greens variety, but if they are they will push the Labor vote out to a lead of around 620 making it very difficult for the Greens to catch up.  However even if they are that strong, they will not put the Liberals close to Labor (about 340 behind).  Labor is also so far beating the Liberals on both absents and dec prepolls.  If the remaining postals are lefter that puts the Greens closer but the Liberals further behind.  Awfully close but I do not see a path in which Labor gets pushed into third.  

Saturday 8:30 A new 3CP is out with votes from extra absents and dec prepolls. Harkin has somehow got ahead now (Harkin 29202 Burns 29152 Hodgins-May 28657).  The AEC say the current vote share is about 93% of expected turnout, which would put expected turnout on 87.5% (down from 89.1%).  To come, about 1450 absents, 300 provisionals, 3000 dec prepolls and maybe 1750 postals and COVID votes.  Burns over Hodgins-May is just 495 votes but unless there are a lot of COVID votes and they are spectacularly Green it's hard to see that being closed.  And if it is, that means COVID votes are left-wing and Burns should get over Harkin again.  Still not completely settled though!

Saturday 9:30 Just ran a 3CP projection on remaining absents and dec prepolls and I have the Greens only about 200 behind (less after provisionals) and the Liberals 130 ahead.  And that will come down still more on provisionals (perhaps to about 140), and then there's remaining postals and provisional votes.  Anything could happen!

Sunday: Nothing has happened today as we prepare for an onslaught of counting tomorrow but the first COVID phone booth has reported in the division of Flinders.  In this case COVID votes break against the Liberals by 11.2 points, to Labor by 3 and to the Greens by 4.2, compared to other ordinary votes.  As applied to Macnamara this would strengthen Labor's chances of just being ahead of both other parties, but I am sceptical that the break to the Greens would be necessarily anything like as weak in Macnamara.  The Green vote is weak in Flinders, partly suppressed by independents.  

Monday: A batch of absents breaks 170-137-114-60 (Greens-Labor-Liberals-others), taking 33 off Labor's lead on primaries and possibly another 10 or so on preferences.  This is not greatly different from other absents so far (a little stronger for Labor and Greens and slightly weaker for the Liberals).  

Monday 2:30: In any other count I wouldn't mention it, but the Liberals just gained nine on rechecking, with Labor and the Greens each down one.  

Monday 3:40: A huge lot of dec prepolls breaks 417-412-404-214 (Greens-Labor-Liberals-others), most likely most of the dent into Labor's lead here would be on preferences.  This is a weaker split (28.8-28,5-27.9) than the split on absents previously (30.8-28-25.6) so the Greens while closing the gap have taken a substantial hit on my projection here  - another issue is that the AJP share of the absents and dec prepolls isn't that high so the flows might be weaker.  It seems now that Labor can only be behind the Greens if there are both a lot of COVID votes and a strong break to the Greens on them (and there is still also the possibility of Labor being over the Liberals.)

Monday 8:45: COVID votes have broken to Labor on primaries (168-135-104 Labor-Greens-Liberal-Other) so that's the end of this epic as a contest and Labor will retain.


  1. Just ran the numbers for myself (Tuesday night) I ended up with Burns 31.96 Hodgins-May 29.26 Harkin 29.64 Other 9.14. Personally I think that gap is slightly too far for Greens, I don't think Hodgins-May can make up enough ground on minor preferences, although I think there is a good chance Harkin makes the final count.

    1. My less sophisticated calculator now has it at Burns 31.79 Hodgins-May 29.53 Harkin 29.49 Other 9.18, with Antony Greens preference estimate that would put it at Harkin 33.90, Burns 33.45, Hodgins-May 32.65. Very tight I can see why its still in doubt. I am probably underestimating Greens support with the later postals.
      fascinating contest.

  2. I've noticed other realignments done by AEC have been in Alphabetical order of booth name, if you want to hazard a guess at which booths are in the Macnamara 3PP

    1. Thankyou. Yes, they usually are more or less alphabetical.

  3. Hey Kevin, firstly thank you for all the great work you do here. I note, in line with your comments above and the 2019 aec data that Labor last election polled ~6% less on postals in this electorate, but that at the moment they are only about 0.5 less. The other parties seem to have values roughly similar to last election. Are you aware of any reason for this? Seems odd and lucky for them considering they're ordinary primary is only up 1.3% and I understood they usually do worse in postals. Any thoughts?

    1. That is interesting. I looked back and found that in 2013 and 2016 Michael Danby did similarly in postals to ordinaries, so 2019 was unusual. Maybe because the seat was vacant.

  4. Luckily for you all I can't post my ghastly spreadsheets here to demonstrate my point/query but, using the preference numbers currently posted on the ABC's Macnamarra page for the 3cp count (49% LIB, 17.7% ALP, 35.4% GRN) in conjunction with a crude projection/extrapolation to the votes listed as "envelopes awaiting processing" on the AEC site of current %'s shown on the AEC site for each of the already partly counted absent, dec prepoll, provisional using dec prepoll % (not sure good number to use here?), and postal with the current 200 + 2000 more it was suggested might come in (what is good here?) then, assuming those AEC % numbers hold steady for categories already partly counted, it seems like the Greens need an increase of about 2.5% and the ALP -2 5% of preferences from the bottom non top 3 parties, in order to just get out in front of ALP in the 3cp.

    Any thoughts on the chances of that minor/micro party preference shift happening in the remaining count, or the Green vote improving in the existing categories? Do we know what booths are awaiting processing etc? Also, any tips on amending some of my assumptions listed above much appreciated.

    What a great 3 way contest!!

    1. I think it will be a lot less than 2000 more postals, maybe a few hundred, maybe several hundred. The thing with them is that late-arriving postals break differently to early arriving postals so much that they can really only be guessed at, not projected off the total to date. The other difficulty is that we have no idea how many COVID votes there are, how they break, or when they will be added. It is possible these are mostly young voters and favourable to the Greens (indeed that is their major hope, since otherwise they appear to be short.)

    2. Ok cheers. With only 200 postals remaining the change required was more like just 1.5% it seemed to me. Which "awaiting processing" category of do the COVID votes fit in to or are they not listed on the AEC site?

    3. There are no booths as such awaiting processing, it is all absents, provisionals, dec prepolls, postals and COVID votes.

    4. Oh ok I see now, there is an 'EAV COVID19 Macnamarra PPVC' polling place, but yeah, as you noted elsewhere it doesn't have any votes listed in it yet. Will keep an eye on it tomorrow or see if it gets noted that they are bundled in one of the declaration vote categories. Cheers.

    5. Flinders now has votes under that polling place.

    6. Ohh yeah, so presumably not included in the Macnamarra tally yet. And with 380 votes there in the Flinders COVID EAV too (hope they all recover well btw). Obviously different dynamics in that seat but looks like about a 10% swing against Lib in the COVID EAV from overall division total, ~3% to ALP and ~4% to Greens with a ~10% swing against Lib on 2pp. Perhaps it will be similar in Macnamarra though Labor seem to be already polling about ~4-5% lower in the other declaration votes (minus postals) there.