VIC and WA federal redistribution drafts released – live

351

6:12pm – I have one last update and then my margins will be finalised.

There are four seats in Victoria where independents made it to the two-candidate-preferred count (2CP), but have added new areas where there was no independent in the 2CP: Goldstein, Kooyong, Nicholls and Wannon.

This issue isn’t relevant in Curtin, since that seat only lost territory. It’s also not an issue in Labor vs Coalition seats with non-classic areas added, since the AEC has calculated a 2PP figure in every part of the country. It’s also not such a big issue in the seat of Melbourne. Since the Greens had a primary vote in the new areas added to Melbourne, you can calculate a margin based on preference flows.

But in the case of Goldstein, Kooyong, Nicholls and Wannon, none of that works. You could theoretically not count any votes in the newly-added areas, or give the independent candidates zero votes in those areas. Neither of those seem fair.

Accounting for these new areas is important in all four seats, but particularly in Kooyong. Almost one quarter of all electors in Kooyong are new to the electorate, all from Higgins. The figure in the other seats ranges from 3.7% in Wannon to 9.1% in Goldstein. This reflects the relatively minor changes in rural Victoria and the major changes in the eastern suburbs of Melbourne.

My first approach earlier this afternoon was to credit all Labor 2PP votes to the independent, and all Coalition votes to the Coalition candidate. But I think that underestimates their support.

In the areas which were not new additions to these seats, we have both a 2CP between the independent and a Liberal or Nationals opponent, plus a 2PP which is Labor vs Coalition. We also have 2PP counts for all the new areas. In all four cases, the newly-added areas are less favourable to Coalition on the 2PP than the areas already contained in these seats. Indeed every seat that gave some territory to a seat where an independent made the final count is held by Labor: Hotham, Isaacs, Higgins, Bendigo and Corangamite.

We know that generally independents did better against the Coalition than Labor did in these seats.

So this table shows my revised approach. I have compared the 2CP and 2PP in the non-moving areas, to calculate how much the independent over-performed Labor. I then add that extra vote to the Labor 2PP in the newly-added areas.

This approach significantly improves the independent position in all four seats. What do you think?

I also want to briefly touch on the peculiar seat of Macnamara. My approach to redistribution (which I believe is similar to Antony and William) is to break up the vote by each SA1, and then reassign the SA1s to the new seats and merge them. Unfortunately this means that, when there is a vote category that has been amalgamated into a single seat-wide total (such as postal votes) effectively I assign the same share of postal votes to every SA1. This is less true for pre-poll votes (where there are multiple pre-poll centres with different geographic patterns) and much less true for election day votes.

I have an alternative approach for state and local redistributions, where we don’t have SA1 results data. For those, I distribute the election day votes then skew the special votes to match the skew of the election day vote. So if Labor does better in one part of the seat on election day, I give it a better share of the special vote in that part of the seat.

I tried to apply that approach to my federal method but it didn’t work, so I’ve left it as is.

Most of the time this doesn’t cause problems. Usually we’re most interested in seats where the changes were significant, not the seats where changes were slight. These estimates are not precise, so when changes are small they should be taken with a grain of salt. 0.1% one way or the other isn’t really meaningful.

Now in Macnamara and Higgins there is a peculiarly large gap between voting patterns in different parts of the seats, and we’ve often seen very left-wing areas around Windsor moved around while they are part of larger seats that have voted Liberal (or at least not been so left-voting). This can produce peculiar outcomes where a small movement of a very left-wing part of a more conservative seat produces a counterintuitive change in the margin.

I recommend that people don’t obsess over very slight changes in the margin or primary vote estimates in Macnamara. The seat was close to a three-way tie in 2022 and any redistribution changes will be much less significant than how voters change in 2025.

4:10pm – I’ve now finished replacing the data after fixing the SA1 issue. The margin in Melbourne has dropped a bit further to 6.9% (I’d previously estimated 7.9%). The Labor margin in Wills is slightly better than I’d previously estimated, now at 4.6%.

3:51pm – Looking at the Victorian 2PP and primary votes, the main changes were Bruce, where the Labor margin is now 5.3%, which is much closer to the pre-redistribution margin and closer to Antony’s margin.

3:06pm – Okay I’ve solved the SA1 problem and will start uploading the corrected figures. Starting with 2PP and primary for WA, the Labor margin for Cowan has dropped to 9.9%, whereas my first estimate had it up to 11.0% (from 10.8%). The Labor margin in Bullwinkel is just 3.3% (not 3.7%) and Labor in Pearce is on 8.8% (not 8.4%). The Liberal margin in Canning is now 1.1%, not 0.8%.

2:33pm – It appears the AEC has switched from using 2016 SA1s for the 2022 election results spreadsheet to 2021 SA1s for the redistribution data, so it will be necessary to add some extra code that adjusts for these changes and this may change some margins. I’ll get that done later today and update the tables.

2:07pm – Okay I’m logging off now. I’m sure there’ll be more analysis later. I will be writing a piece for the Guardian tomorrow and I’ll be carefully kicking some tyres to see if there are any errors in the estimates over the coming days.

If you appreciated this very quick analysis of the breaking news, please consider signing up to support The Tally Room on Patreon!

2:05pm – So the creation of Bullwinkel in the outer east of Perth has then pushed all of the neighbouring seats out of the way.

Hasluck has become a much smaller seat and now sits entirely on the northern side of Perth.

Moore has shifted south, with Pearce adding a small area from Moore. Cowan and Perth have lost their eastern edges to Hasluck.

Swan has also shifted west, while Canning has lost its north-eastern corner to Bullwinkel and compensated by picking up Karnup from Brand. This explains the big drop in the Liberal margin there.

Tangney, Fremantle and Curtin have experienced very minor changes.

1:59pm – And here is the interactive map for WA.

1:55pm – Okay now here we have the 2CP margins for WA. Curtin thankfully didn’t add any extra territory so no complicated calculations needed there. Bullwinkel is a notional Labor seat with a 3.7% margin. Labor has also significantly improved their position in Hasluck, and the Liberal position is weaker in Canning. Labor’s margin has also been dented in Burt.

This means that Labor has gained a seat and the Liberal Party has lost a seat so far in this redistribution, with NSW yet to come.

Seat Old margin New margin
Brand ALP 16.7% ALP 17.1%
Bullwinkel (new) ALP 3.3%
Burt ALP 15.2% ALP 13.3%
Canning LIB 3.6% LIB 1.1%
Cowan ALP 10.8% ALP 9.9%
Curtin IND vs LIB 1.3% IND vs LIB 1.3%
Durack LIB 4.3% LIB 4.7%
Forrest LIB 4.3% LIB 4.2%
Fremantle ALP 16.9% ALP 16.7%
Hasluck ALP 6% ALP 10.1%
Moore LIB 0.7% LIB 0.9%
O’Connor LIB 7% LIB 6.7%
Pearce ALP 9% ALP 8.8%
Perth ALP 14.8% ALP 14.4%
Swan ALP 8.8% ALP 9.4%
Tangney ALP 2.4% ALP 3%

1:49pm – And here we go with WA. This table shows the 2PP and primary vote estimates for each seat.

Seat ALP 2PP LIB 2PP ALP prim LNP prim GRN prim IND prim
Brand 67.1 32.9 50.7 21.8 11.3 0.0
Bullwinkel 53.3 46.7 36.4 35.74 11.3 1.8
Burt 63.3 36.7 49.8 24.78 9.5 0.2
Canning 48.9 51.1 35.1 41.44 8.4 1.6
Cowan 59.9 40.1 45.8 30.97 10.0 0.0
Curtin 44.4 55.6 13.8 41.36 10.4 29.7
Durack 45.3 54.7 28.8 44.84 9.5 0.0
Forrest 45.8 54.2 27.7 43.13 13.3 0.1
Fremantle 66.7 33.3 44.0 24.38 17.9 0.0
Hasluck 60.1 39.9 43.7 30.12 11.4 2.1
Moore 49.1 50.9 31.9 41.81 14.1 1.3
O’Connor 43.3 56.7 26.7 44.5 10.9 0.0
Pearce 58.8 41.2 42.4 30.12 11.2 0.0
Perth 64.4 35.6 39.1 27.21 22.0 0.0
Swan 59.4 40.6 40.0 31.64 15.1 0.0
Tangney 53.0 47.0 38.2 39.41 12.4 0.0

1:36pm – So just a quick description of what the map shows before moving on to WA.

The seat of Melbourne has jumped the river into South Yarra, which has then pulled Wills and Cooper south, making Wills much stronger for the Greens. This doesn’t appear to have done much to the Greens’ position in Macnamara, although we’ll need to wait for a 3CP estimate to know for sure.

The abolition of Higgins has had dramatic impacts in the eastern suburbs, with Kooyong and Chisholm absorbing most of the seat.

Menzies has lost areas further east and expanded into Box Hill, which explains the seat becoming notional Labor.

Deakin has retracted to areas further east, further reducing the Liberal margin from a slim 0.2% to 0.02%.

Aston was barely touched, as was Goldstein, but Hotham, Isaacs and Dunkley have all been pulled north. Casey also expanded west to take in areas from McEwen and Menzies.

In the western suburbs, Lalor has contracted sharply, and Hawke has taken in the area around Melbourne Airport. But generally changes in the west were mild.

Outside of Melbourne, Corangamite has again shrunk in size, now almost entirely fitting within the Bellarine Peninsula.

Indi, Gippsland, Mallee and Monash appear to be unchanged, or close to it. McEwen has moved closer to Melbourne, but it has made no difference to the margin.

1:26pm – Okay I have now had a chance to revise my Melbourne 2CP estimate which was very quick. I now have the Greens on 7.9% by applying the same preference flows to the new areas as the rest. That is a drop in the Greens margin of 2.3%, but nothing like my first estimate.

1:18pm – Okay here is my interactive map where you can toggle between the old and new boundaries for Victoria. Will take a quick bathroom break then be back.

1:05pm – And here is my first stab at the new margins for Victorian seats compared to the old margins.

Seat Old margin New margin
Aston LIB 2.8% LIB 2.6%
Ballarat ALP 13% ALP 13%
Bendigo ALP 12.1% ALP 12%
Bruce ALP 6.6% ALP 5.3%
Calwell ALP 12.4% ALP 12.4%
Casey LIB 1.5% LIB 1.4%
Chisholm ALP 6.4% ALP 3.3%
Cooper ALP vs GRN 8.7% ALP vs GRN 7.8%
Corangamite ALP 7.6% ALP 7.8%
Corio ALP 12.8% ALP 12.5%
Deakin LIB 0.2% LIB 0%
Dunkley ALP 6.3% ALP 6.8%
Flinders LIB 6.7% LIB 6.2%
Fraser ALP 16.5% ALP 16.6%
Gellibrand ALP 11.5% ALP 11.2%
Gippsland NAT 20.6% NAT 20.6%
Goldstein IND vs LIB 2.9% IND vs LIB 3.9%
Gorton ALP 10% ALP 10%
Hawke ALP 7.6% ALP 7.6%
Higgins (abolished) ALP 2.1%
Holt ALP 7.1% ALP 7.1%
Hotham ALP 14.3% ALP 11.6%
Indi IND vs LIB 8.9% IND vs LIB 8.9%
Isaacs ALP 6.9% ALP 9.5%
Jagajaga ALP 12.3% ALP 12.2%
Kooyong IND vs LIB 2.9% IND vs LIB 3.5%
La Trobe LIB 8.7% LIB 8.4%
Lalor ALP 12.8% ALP 12.8%
Macnamara ALP 12.2% ALP 12.2%
Mallee NAT 19% NAT 19%
Maribyrnong ALP 12.4% ALP 13%
McEwen ALP 3.3% ALP 3.4%
Melbourne GRN vs ALP 10.2% GRN vs ALP 6.9%
Menzies LIB 0.7% ALP 0.4%
Monash LIB 2.9% LIB 2.9%
Nicholls NAT vs IND 3.8% NAT vs IND 2.5%
Scullin ALP 15.6% ALP 15.3%
Wannon LIB vs IND 3.9% LIB vs IND 3.4%
Wills ALP vs GRN 8.6% ALP vs GRN 4.6%

The Greens margin in Melbourne has been weakened quite significantly, while the Greens are much closer in Wills. Labor is also slightly weaker in Cooper.

Labor is much weaker in Bruce, Chisholm and Hotham, but stronger in Isaacs.

The seat of Menzies has flipped from 0.7% for the Liberal Party to 0.3% for Labor.

With Labor losing Higgins but picking up Menzies, that’s a net loss of one seat for the Liberal Party.

12:47pm – Okay I have calculated the 2PP and primary vote for the main parties for each seat, below.

Seat ALP 2PP LNP 2PP ALP prim LNP prim GRN prim IND prim
Aston 47.4 52.6 32.5 42.8 12.2 0.1
Ballarat 63.0 37.0 44.8 27.1 14.5 2.1
Bendigo 62.0 38.1 42.8 26.7 14.0 4.4
Bruce 55.3 44.7 40.3 31.7 9.7 0.2
Calwell 62.4 37.6 44.9 23.7 9.8 0.0
Casey 48.6 51.4 25.1 36.6 13.1 11.4
Chisholm 53.3 46.7 35.0 39.2 13.8 4.0
Cooper 75.7 24.3 40.7 16.2 28.4 0.0
Corangamite 57.8 42.2 38.4 34.0 15.3 0.0
Corio 62.5 37.5 41.9 25.0 14.7 0.1
Deakin 50.0 50.0 32.9 41.5 14.2 1.1
Dunkley 56.8 43.2 40.5 31.7 10.6 3.4
Flinders 43.8 56.2 22.8 43.3 9.5 11.7
Fraser 66.6 33.4 42.1 24.5 18.9 0.0
Gellibrand 61.2 38.8 42.8 27.2 15.6 0.3
Gippsland 29.4 70.6 19.2 54.1 8.5 0.0
Goldstein 46.3 53.7 13.6 39.6 8.4 31.3
Gorton 60.0 40.0 41.3 27.4 9.0 2.5
Hawke 57.6 42.4 36.7 26.4 8.9 7.9
Holt 57.1 42.9 40.8 29.5 8.6 3.0
Hotham 61.6 38.4 42.9 28.6 15.0 0.2
Indi 44.7 55.3 8.6 34.3 3.6 40.7
Isaacs 59.5 40.5 42.8 29.5 12.1 0.0
Jagajaga 62.2 37.8 40.8 29.2 16.7 3.0
Kooyong 46.3 53.7 11.3 43.4 9.9 31.0
La Trobe 41.6 58.4 26.2 45.2 10.9 0.0
Lalor 62.8 37.2 44.1 25.0 10.4 2.8
Macnamara 62.2 37.8 31.7 29.1 29.7 1.9
Mallee 31.0 69.0 16.8 49.1 5.3 12.2
Maribyrnong 63.0 37.0 42.2 26.8 16.7 0.0
McEwen 53.4 46.6 36.9 33.1 14.2 0.0
Melbourne 73.1 26.9 25.7 19.5 44.7 1.0
Menzies 50.4 49.6 31.8 41.0 12.9 4.9
Monash 47.1 52.9 25.6 37.8 9.9 10.7
Nicholls 34.1 65.9 13.2 43.5 3.7 24.0
Scullin 65.3 34.7 46.1 21.9 10.9 0.0
Wannon 41.4 58.6 19.7 44.2 6.7 20.8
Wills 77.1 22.9 36.4 16.2 32.8 0.2

12:40pm – The AEC has now published the Victorian redistribution. I’m going to focus on getting the new margins up first then analyse the trends.

12:17pm – While the AEC has not published anything, the Gazettes are now up.

In Victoria, the seat of Higgins has been proposed to be abolished. No other seat has changed names, and apparently 34 other divisions have been changed. 8.31% of all electors have been moved to a new seat.

In Western Australia, the new seat is named Bullwinkel, after Lieutenant Colonel Vivian Bullwinkel. The seat seems to be located in the outer eastern suburbs of Perth. 14.57% of electors have been moved to new seats.

12:00pm – The Australian Electoral Commission will be announcing the draft federal electorate boundaries for the states of Western Australia and Victoria this afternoon. They have indicated that the boundaries will be published at some point between 12:30pm and 2:30pm AEST.

My plan is to publish my estimated margins for each electorate, and estimated primary votes for the main party groupings, some descriptions of what changes have happened, and maps showing the old and new boundaries.

In 2021 I was held up by a problem where they didn’t publish the SA1s for Victoria until a couple of hours after they published their report, and then there was a problem with the data. Hopefully that won’t happen again, but I’ll be relying on that data to calculate the new margins.

On the other hand, I have previously drawn my own KML versions of the electorate boundaries. I am not planning to do that this time, so it should be quicker to take the AEC shapefile and make interactive maps this afternoon.

Liked it? Take a second to support the Tally Room on Patreon!
Become a patron at Patreon!

351 COMMENTS

LEAVE A REPLY

Please enter your comment!
Please enter your name here