The Victorian Electoral Commission has recently published a report laying out their estimates of the two-candidate-preferred margins in every electorate for the upcoming Victorian state election, after taking account of the redistribution.
I’ve previously published my own estimates, but I’m planning to substitute the VEC’s margins for my own. But before I do that I thought it would be interesting to compare the two estimates – while they aren’t the same, they are pretty similar.
The Australian Electoral Commission consistently produces redistribution estimates and it sometimes happens in other states, but not always. Sometimes we’re entirely reliant on non-official publications from people like myself, William Bowe or Antony Green. And even when they are published, they usually come out at a much slower pace.
But one of the interesting things about this report is that the VEC explain in some detail their methodology, including providing the detail for one electorate as an example.
The process has two steps. For ordinary election day votes, they use (non-public) data on what share of each booth’s voters have been moved into each seat to divide up the booth’s votes in even proportions for each party. For special votes, they split them up in proportion, but they skew the partisan share to reflect the partisan skew amongst the ordinary votes. So if Labor has done better in the ordinary vote for one part of the seat, they do better for the special vote in that area too.
This is quite similar to the method I use for redistributing state and local elections. I don’t have access to the data about how many people from each booth are in each geographic area for Victorian state elections. That data is only published for federal elections. So I transfer whole polling places, and where the proportions are way off I sometimes split a border booth in half or some other fraction. But the rest of the method is pretty similar.
At the end of this post I’ve published the pre-redistribution margins and both my and the VEC’s post-redistribution margins for all seats, but I’ve summarised some of the interesting changes.
The VEC has flipped the party holding two seats, compared to my own estimates. They have the Nationals 0.7% better than I estimated in Mildura, which is enough to flip it from a very marginal independent seat to a very marginal Nationals seat. They have also flipped Caulfield from being very marginal Liberal (0.1% on my estimate) to very marginal Labor (0.2%).
The 2CP pairings changed. I had included the independent in the margin in Geelong and Pascoe Vale, while the VEC included the Liberal. But in both cases I had also found a 2PP margin that was no more than 0.1% off. In the seat of Morwell, the incumbent independent (who is retiring) was 1.1% behind Labor on my estimate, but the VEC has produced a 4% margin for Labor against the Nationals.
Overall the margins are very similar. Comparing all 88 seats (including the 2PP margins where the pairings differed), 29 seats have exactly the same margin, 60 are within 0.2%, 74 are within 0.5%, and 88 are within 1.0%. The biggest divergence is in Monbulk, where my estimate was 9.0% while the VEC has 7.1%.
That’s it for now, but I will be making updates to the margins in the guide in the coming days.
| Seat | Old margin | New margin (BR) | New margin (VEC) |
| Albert Park | ALP 13.1% | ALP 12.9% | ALP 13.1% |
| Ashwood (Burwood) | ALP 3.3% | ALP 2.3% | ALP 2.0% |
| Bass | ALP 2.4% | LIB 0.7% | LIB 0.7% |
| Bayswater | ALP 0.4% | LIB 0.7% | LIB 0.6% |
| Bellarine | ALP 11.5% | ALP 11.4% | ALP 11.4% |
| Benambra | LIB 2.4% | LIB vs IND 2.6% | LIB vs IND 2.6% |
| Bendigo East | ALP 12.1% | ALP 12.1% | ALP 12.1% |
| Bendigo West | ALP 18.6% | ALP 18.6% | ALP 18.6% |
| Bentleigh | ALP 11.9% | ALP 11.3% | ALP 11.4% |
| Berwick (Gembrook) | LIB 0.8% | LIB 2.1% | LIB 1.3% |
| Box Hill | ALP 2.1% | ALP 2.8% | ALP 3.1% |
| Brighton | LIB 1.1% | LIB 0.5% | LIB 0.5% |
| Broadmeadows | ALP 30.3% | ALP 24.3% | ALP 25.2% |
| Brunswick | GRN vs ALP 0.6% | GRN vs ALP 2.0% | GRN vs ALP 2.0% |
| Bulleen | LIB 5.8% | LIB 5.7% | LIB 5.5% |
| Bundoora | ALP 17.4% | ALP 16.0% | ALP 16.2% |
| Carrum | ALP 11.9% | ALP 12.2% | ALP 12.0% |
| Caulfield | LIB 0.3% | LIB 0.1% | ALP 0.2% |
| Clarinda | ALP 17.4% | ALP 14.9% | ALP 14.9% |
| Cranbourne | ALP 11.0% | ALP 9.1% | ALP 9.3% |
| Croydon | LIB 2.1% | LIB 0.9% | LIB 1.0% |
| Dandenong | ALP 23.9% | ALP 23.5% | ALP 23.1% |
| Eildon | LIB 2.4% | LIB 1.0% | LIB 1.0% |
| Eltham | ALP 9.1% | ALP 8.8% | ALP 8.8% |
| Essendon | ALP 15.9% | ALP 15.8% | ALP 15.8% |
| Eureka (Buninyong) | ALP 12.2% | ALP 9.7% | ALP 9.6% |
| Euroa | NAT 15.4% | NAT 16.0% | NAT 15.8% |
| Evelyn | LIB 2.6% | LIB 1.6% | LIB 1.8% |
| Ferntree Gully | LIB 1.6% | Abolished seat | Abolished seat |
| Footscray | ALP 28.1% | ALP 29.1% | ALP 27.6% |
| Frankston | ALP 9.7% | ALP 10.1% | ALP 10.2% |
| Geelong | ALP vs IND 6.2% | ALP vs IND 6.1% | ALP 10.3% |
| Gippsland East | NAT 17.6% | NAT 17.6% | NAT 17.6% |
| Gippsland South | NAT 15.3% | NAT 13.7% | NAT 14.2% |
| Glen Waverley (Forest Hill) | LIB 1.2% | LIB 1.5% | LIB 0.9% |
| Greenvale | New seat | ALP 22.7% | ALP 22.0% |
| Hastings | LIB 1.1% | ALP 0.4% | ALP 0.01% |
| Hawthorn | ALP 0.4% | ALP 0.5% | ALP 0.6% |
| Ivanhoe | ALP 12.4% | ALP 12.7% | ALP 12.3% |
| Kalkallo (Yuroke) | ALP 20.3% | ALP 20.1% | ALP 20.9% |
| Kew | LIB 4.8% | LIB 4.8% | LIB 4.7% |
| Keysborough | ALP 14.9% | Abolished seat | Abolished seat |
| Kororoit | ALP 25.6% | ALP 25.3% | ALP 25.3% |
| Lara | ALP 19.1% | ALP 19.1% | ALP 19.1% |
| Laverton | New seat | ALP 23.4% | ALP 23.4% |
| Lowan | NAT 23.5% | NAT 20.9% | NAT 21.1% |
| Macedon | ALP 13.2% | ALP 13.4% | ALP 13.4% |
| Malvern | LIB 6.1% | LIB 6.6% | LIB 6.0% |
| Melbourne | GRN vs ALP 1.3% | GRN vs ALP 1.7% | GRN vs ALP 1.7% |
| Melton | ALP 4.3% | ALP 5.4% | ALP 5.0% |
| Mildura | IND vs NAT 0.3% | IND vs NAT 0.3% | NAT vs IND 0.4% |
| Mill Park | ALP 24.9% | ALP 24.9% | ALP 24.9% |
| Monbulk | ALP 8.6% | ALP 9.0% | ALP 7.1% |
| Mordialloc | ALP 12.9% | ALP 13.5% | ALP 13.4% |
| Mornington | LIB 5.0% | LIB 5% | LIB 5.0% |
| Morwell | IND vs ALP 1.8% | ALP vs IND 1.1% | ALP 4.0% |
| Mount Waverley | ALP 1.8% | Abolished seat | Abolished seat |
| Mulgrave | ALP 12.7% | ALP 16.2% | ALP 15.8% |
| Murray Plains | NAT 23.9% | NAT 24.0% | NAT 24.0% |
| Narracan | LIB 7.3% | LIB 11.0% | LIB 10.0% |
| Narre Warren North | ALP 9.8% | ALP 10.2% | ALP 10.4% |
| Narre Warren South | ALP 6.9% | ALP 10.7% | ALP 10.4% |
| Nepean | ALP 0.9% | ALP 0.6% | ALP 0.7% |
| Niddrie | ALP 12.6% | ALP 12.7% | ALP 12.5% |
| Northcote | ALP vs GRN 1.7% | ALP vs GRN 1.7% | ALP vs GRN 1.7% |
| Oakleigh | ALP 15.8% | ALP 16.1% | ALP 16.0% |
| Ovens Valley | NAT 12.6% | NAT 12.0% | NAT 12.1% |
| Pakenham | New seat | ALP 2.0% | ALP 2.2% |
| Pascoe Vale | ALP 8.3% | ALP vs IND 9.1% | ALP 22.3% |
| Point Cook (Altona) | ALP 14.6% | ALP 12.3% | ALP 12.8% |
| Polwarth | LIB 5.4% | LIB 2.5% | LIB 2.0% |
| Prahran | GRN vs LIB 7.5% | GRN 9.4% | GRN 8.2% |
| Preston | ALP vs GRN 20.7% | ALP vs GRN 21.2% | ALP vs GRN 21.3% |
| Richmond | ALP vs GRN 5.5% | ALP vs GRN 5.8% | ALP vs GRN 5.8% |
| Ringwood | ALP 2.8% | ALP 3.7% | ALP 3.2% |
| Ripon | LIB 0.02% | ALP 2.8% | ALP 2.8% |
| Rowville | LIB 5.7% | LIB 5.3% | LIB 5.5% |
| Sandringham | LIB 0.6% | LIB 0.4% | LIB 0.4% |
| Shepparton | IND vs LIB 5.3% | IND vs LIB 5.3% | IND vs LIB 5.3% |
| South Barwon | ALP 4.6% | ALP 3.7% | ALP 3.0% |
| South-West Coast | LIB 2.3% | LIB 3.3% | LIB 3.2% |
| St Albans | ALP 21.5% | ALP 21.9% | ALP 22.0% |
| Sunbury | ALP 14.3% | ALP 14.6% | ALP 14.5% |
| Sydenham | ALP 17.9% | ALP 17.9% | ALP 18.3% |
| Tarneit | ALP 18.0% | ALP 17.9% | ALP 17.9% |
| Thomastown | ALP 27.2% | ALP 27.3% | ALP 27.4% |
| Warrandyte | LIB 3.9% | LIB 3.9% | LIB 3.8% |
| Wendouree | ALP 10.3% | ALP 11.0% | ALP 11.0% |
| Werribee | ALP 12.6% | ALP vs IND 9.2% | ALP vs IND 9.1% |
| Williamstown | ALP 22.1% | ALP 18.7% | ALP 19.9% |
| Yan Yean | ALP 17.0% | ALP 16.9% | ALP 16.9% |
Hi Ben, just a correction, the VEC press release actually listed Caulfield as flipping to notional Labor on a 0.2% margin, not 0.2% LIB.
Interesting how the VEC’s estimate has Malvern going the away after it gained the balance of Toorak. I suppose the western end of Toorak does have more apartments and fewer mansions.
@Nicholas The Greens and Labor also put more effort into the Prahran part of Toorak compared to the rest of it in Malvern which could lower the Lib vote there.
Nicholas I was curious about that too.
But I think what it might come down to is just that it’s very specific to the 2018. I think in normal circumstances, gaining the balance of Toorak would increase the Liberal margin but those two booths happened to have something like 18% swings in 2018 and one of them for example only had a 51% Liberal 2PP which is actually less than Malvern’s 2PP.
So my view is it’s probably more of an anomaly due to the wild 2018 swings but in reality it will actually increase the Liberal margin a bit.
Conversely, I think that’s also why the Greens’ Prahran margin only went up to 8.2% – the impact of the Toorak booths being moved out probably had less impact due to the big 2018 swings than it usually would when removing a larger Liberal vote would probably increase the Greens margin to over 9%.
Thanks Trent, will fix Caulfield.
Caulfield has never been held by the ALP & a target of theres, the LNP have a massive hill to climb if they want to gain any ground this election.
Nicholas – Toorak has been slowly shifting away from the Liberals. However, its too soon to tell if its a permanent change or simply because the Liberals have been performing poorly on key issues. Toorak is as posh and rich as ever and for that reason its more about the Liberals.
Footscray is likely an ALP versus Greens seat, even on the old boundaries, as their preferences are unlikely to favour the Greens by less than the 1.8% primary vote gap. The new boundaries are quite favourable to the Greens in Footscray, at the expense of the Greens in Williamstown. The redivision adjusted ALP versus Green margin in Footscray is almost certainly considerably less than the ALP versus Liberal.
Pascoe Vale also has a reasonable chance at an ALP versus Greens 2CP, although to be fair that is more subjective than Preston or Footscray because of the significant independent vote last time.
Toorak is less friendly to the centre-left/teal votes due to the ultra-rich millionaires living in mega-mansions. I think it seems Toorak is more libertarian/IPA types rather than small-Liberals (they may be slightly conservative economic such as globals investments and economic growth but support some egalitarian economic policies with some government intervention)
I find it interesting that the VEC has the redistribution in Monbulk dropping Labor’s margin. Ferntree Gully in the west and Emerald in the east are the main additions, and I certainly would’ve thought that those areas are stronger for Labor than Montrose, Kilsyth and Mooroolbark which they replaced. Nick Waekling in Ferntree Gully and Brad Battin in Gembrook would also have had some personal vote in those areas, which will now be gone, although of course Merlino had a huge personal vote which is gone. I think ALP will still retain by a couple of % points.
Agree, Adam with your analysis. I feel there maybe swings in different directions in Monbulk depending on the loss of personal vote of the previous representative.
Oscar Yildiz’s Victorians Party have announced they are no longer contesting the election. They seemed to have disappeared completely with their website no longer active and social media accounts deleted.
This is going to be a very sad election.
Comments are closed.