Following the publication of the proposals for the Victorian and WA redistributions, I put together two posts analysing different angles at a deeper level than the original analysis.
Rather than fully revisiting the topics, I thought I would just use one post to apply the same analysis to the NSW proposal, and also zoom out to the national level.
The two posts looked at how many seats were in each bracket of marginality prior to and after the redistribution, and how the old and new boundaries compare looking at results from 2004 to 2022.
Firstly, I’ve produced these charts that show how many seats have a particular margin on the two-party-preferred vote in the three redrawn states.
Coalition seats in NSW have on average become less marginal. There is one less seat in the 60-65% range and one less seat in the 55-60% range. These seats haven’t been abolished, but have instead appeared in the 50-55% range. Prior to the redistribution, there were three seats with a Coalition 2PP in the 50-55% range, two of which were won by independents – Warringah and North Sydney. After the redistribution, there were five. Again, two of them are held by independents. North Sydney was abolished, but Wentworth became stronger for Labor on the 2PP and entered that range.
Labor lost one seat, which is reflected in their reduced number in the 50-55% range. Their remaining seats are a bit less safe, with one seat moving from 60-65% to 55-60% range.
Now that we have redistributions for the three big states, I’ve also done a similar chart for the whole country.
Coalition seats generally became less safe, with one seat moving from the 60-65% range to the 50-55% range. The same is also true on the Labor side, with one less seat at 65-70% and one more at 50-55%.
Next up, I’ve converted the results from all federal elections from 2004 to 2022 to the new proposed boundaries for NSW. At the end of the post I’ve listed every NSW seat with the actual 2019 result and the 2019 result for the new boundaries.
For this next chart, I show my estimate of the Labor 2PP in a selection of seats both for the 2022 version of electorates, and the 2025 version. I’ve now also added a third line to show the actual outcomes at each election.
These aren’t the most intuitive graphs but when you pay attention they have an interesting story. When the blue and red lines have diverged, that indicates a previous change in a seat’s political make-up. When the green line diverges from the other lines, that indicates a change between 2022 and 2025.
The proposed boundaries in Barton are consistently worse for Labor than the 2022 boundaries, it’s not as favourable to the Liberal Party as the previous boundaries used from 2004 (or earlier) until 2013, which led to the Liberal Party winning narrowly in 2013. On the proposed boundaries, Labor would have held on in that year.
Bennelong has experienced no boundary changes since 2006, and the 2001-2004 boundaries were only subtly different, so you see no difference on my chart between the actual result and the 2022 boundaries result. It’s interesting to see how the proposed new boundaries would have always been more favourable for the Liberal Party, but the size of that gap has varied over time. It was biggest in 2010, when an incumbent Labor MP was up for election in the actual Bennelong but obviously not in the newly-added suburbs. Perhaps an encouraging sign for the local member.
Greenway’s 2022 boundaries have produced a very similar outcome as all boundaries since 2010, but the 2007 boundaries were dramatically different, with the entire Hawkesbury region added to the seat. The 2004 boundaries covered most of the current area, plus some more of the Blacktown LGA, and thus produced similar results. The proposed changes would have made the Liberal position about 4-5% stronger at any point in the last 20 years compared to the 2022 boundaries.
The new boundaries for Hughes would have narrowly elected Labor in 2007. The last redistribution in 2016 made Hughes slightly better for the Liberal Party, but the proposed boundaries have a much bigger impact in the opposite direction.
It is amazing to see how little difference there is in the seat of Hume despite the seat changing so much, both compared to its 2022 boundaries and its boundaries in earlier eras. The only exception is in 2013, when the ALP would have done much worse in the new Macarthur-based seat than the current boundaries, and even that was worse than the actual outcome in 2013, when the seat was much less urban than the 2022 version.
The gap between the Parramatta result on the old and new boundaries is relatively consistent, although the gap was much smaller for the 2022 results. This would likely reflect the Labor vote being relatively stronger in Bennelong in 2022 than in previous elections, when compared to Parramatta. It may also reflect that Andrew Charlton did relatively better in the north-east of Parramatta and less well in some of the areas lost from the seat. Interestingly, the proposed Parramatta looks a lot similar to the 2004-2007 version when Julie Owens was first winning, but the 2009 redistribution strengthened Labor’s position.
It looks like the gap between the Riverina margin on the 2022 and proposed 2025 boundaries has been widening.
Likewise the gap in Wentworth has been widening, although the 2007-2013 versions of Wentworth were about halfway between the 2022 and 2025 versions.
Finally Whitlam is an interesting seat, with the proposed boundaries consistently worse than the 2022 boundaries, but the shift is much less dramatic than what we saw in 2009 when the former seat of Throsby was first expanded into the Southern Highlands.
In the process of making this chart I also developed a second chart which shows the gap between each metric, rather than the total 2PP. This chart shows 16 seats in all three redistributed states: blue shows the gap between the proposed 2025 boundaries and the actual result, and red shows the gap between the proposed 2025 and 2022 boundaries.
This shows up some fascinating trends more clearly, such as:
- Bennelong – Labor is disadvantaged by the changed boundaries, but would have fared much worse if this change was made prior to 2016.
- Chisholm – the Labor disadvantage from the last redistribution would have been worth about 3-4%, but that’s only about half as much Labor disadvantage they’ve accumulated since the 2007-2013 version.
- Dunkley – the recent change was quite subtle, although it would’ve been worse in past years, and is much less significant than the 2018 change.
- Hotham – Labor’s position has been made about 3-5% worse, but the Labor position is now similar to their actual result prior to the 2021 redistribution.
- Hume – This change would’ve helped the Liberal Party at any point up to 2013, but is now very slightly helpful to Labor.
- Menzies – While this seat flipped from Liberal to Labor, the change was much more subtle than previous changes.
- Wentworth – The Labor 2PP is boosted by about 5%, and it would’ve been similar on the 2016 or 2019 results, but before that election it would’ve only been worth about 3%.
And finally, as promised here is the full list of 2019 results by proposed seat.
| Seat | 2019-2022 map | 2025 map |
| Banks | LIB 6.3% | LIB 5.7% |
| Barton | ALP 9.4% | ALP 6.1% |
| Bennelong | LIB 6.9% | LIB 8% |
| Berowra | LIB 15.7% | LIB 13.8% |
| Blaxland | ALP 14.7% | ALP 12% |
| Bradfield | LIB 16.6% | LIB 15.7% |
| Calare | NAT 13.3% | NAT 13.3% |
| Chifley | ALP 12.4% | ALP 12.7% |
| Cook | LIB 19% | LIB 17.6% |
| Cowper | NAT 11.9% | NAT 11.9% |
| Cunningham | ALP 13.4% | ALP 14.1% |
| Dobell | ALP 1.5% | ALP 1.5% |
| Eden-Monaro | ALP 0.8% | LIB 0.7% |
| Farrer | LIB 19.8% | LIB 19.8% |
| Fowler | ALP 14% | ALP 13.6% |
| Gilmore | ALP 2.6% | ALP 2.6% |
| Grayndler | ALP 23.8% | ALP 22.2% |
| Greenway | ALP 2.8% | LIB 1% |
| Hughes | LIB 9.9% | LIB 5.8% |
| Hume | LIB 12.9% | LIB 12.6% |
| Hunter | ALP 3% | ALP 4% |
| Kingsford Smith | ALP 8.8% | ALP 7.7% |
| Lindsay | LIB 5.1% | LIB 4.7% |
| Lyne | NAT 15.2% | NAT 15.2% |
| Macarthur | ALP 8.4% | ALP 10.2% |
| Mackellar | LIB 13.2% | LIB 12.2% |
| Macquarie | ALP 0.2% | LIB 0.6% |
| McMahon | ALP 6.6% | ALP 7.3% |
| Mitchell | LIB 18.6% | LIB 18.4% |
| New England | NAT 17.6% | NAT 16.3% |
| Newcastle | ALP 13.8% | ALP 13.8% |
| North Sydney | LIB 9.3% | |
| Page | NAT 9.4% | NAT 9.4% |
| Parkes | NAT 16.9% | NAT 17.9% |
| Parramatta | ALP 3.5% | ALP 0.8% |
| Paterson | ALP 5% | ALP 4.2% |
| Reid | LIB 3.1% | LIB 3.1% |
| Richmond | ALP 4.1% | ALP 4% |
| Riverina | NAT 19.5% | NAT 15.1% |
| Robertson | LIB 4.2% | LIB 4.2% |
| Shortland | ALP 4.5% | ALP 4.7% |
| Sydney | ALP 18.7% | ALP 18.8% |
| Warringah | LIB 2.1% | LIB 3.4% |
| Watson | ALP 13.5% | ALP 14.9% |
| Wentworth | LIB 9.9% | LIB 5.6% |
| Werriwa | ALP 5.5% | ALP 5.3% |
| Whitlam | ALP 10.9% | ALP 8.3% |
Good work Ben! I find these super interesting
So the Libs would of won +2 in 2019
+ Macquarie
+ Eden Monaro
+ Greenway
– North Sydney (abolished)
Interestingly Labor wouldn’t of won Bennelong in 2007, and seems to have swung a bit to Labor since 2022. This is different than the old boundaries where Labor did .4% better in Bennelong in 2007 vs 2022
Greenway barely being a Labor win even in 2007! That area has swung massively to Labor. It’s quite bizarre how safe on paper that seat looks, when those boundaries are pretty bad for Labor. Labor would be hoping this was part of a much bigger trend and not just a one-off swing.
Labor winning in Hughes in 2007 matches my initial thoughts that this seat was winnable for them in the right circumstances. I think Labor would have put a lot more effort here in 2022 if it had these boundaries going into the election.
I thought Libs would have won Parramatta by more in 2013.
One issue that is very hard to codify is when an area is moved to or from a safe seat to either a marginal seat or a seat that belongs to the opposing party. It is pretty clear that the major parties put more effort and resources into the seats they hold or marginal seats (their own or the opposing teams). Therefore on the current redistribution it could be said that the new NP margin in Riverina is understated as the areas gained in Snowy Valleys Shire did well for Labor last time but there have been big swings (15% or so) each time the area was moved from Eden Monaro (marginal) to Riverina (safe NP). As that switch has happened a few times, you can see the pattern – Labor don’t put resources in when they know they can’t win. On this basis, it could be said that Hughes is mildly understated for the Libs as they would never have put resources into Glenfield before, something similar could also be said for Whitlam as it is now could be said to be in a potential winnable range especially also as the coastal areas did swing to the Libs in 2022.
@Drake its interesting because even in 2013 when the Coalition won a landslide which was particularly seen in NSW where the Coalition gained 10 seats and Labor lost eight, Labor somehow managed to retain many federal seats in Western Sydney where the Coalition won overlapping state seats and held on to many of them until 2023. Very competitive state seats that are located within federal Labor seats that despite their margins haven’t been nearly as competitive.
For example, Labor marginally retained the seats of Greenway and Parramatta in 2013 even though the Liberals held Parramatta and Riverstone from 2011 to 2023, plus they still hold Winston Hills.
This is the chance the Liberals have to win or marginalise Greenway.
It’s historically been marginal: 2022 for the Libs was a disaster. They ran a sluggish campaign, and that meant blowout swings against them in areas north of the M7, especially those new housing estates which traditionally lean Liberal but collapsed in support. If they get a strong candidate, strong policies which benefit those in Greenway, and can capitalise on Labor and Michelle Rowland making a potential gaffe or just being generally unpopular, then the sky’s is the limit.
Then again, taking down a minister would be huge, and she has massive personal vote, so the Libs can certainly reduce the margin, but a win is possible but very difficult.
the one thing noones talking about or measuring is what would a chisholm-higgins and warringah-wentworth look like margin wise?
obviously in chisholm-higgins youd favour labor but a warringah-wentworth would be interesting. who would labor and the minors preference? stegall or spender? would we get a spender vs stegall contest? could they split in the vote deliver it back to the libs? who would the libs preference if it was Spender vs Stegall?
@james i dont think its marginal enough to win this time around but i think they can get it within striking distance for 2028 along with Macquarie.
@John I doubt it would be a Steggall vs Spender contest.
@np good point the combined lnp vote would probably put them in the 2cp.amd it would come down to preferences at which point it would be a contest on who Labor preference and would provably be directed towards stegall knowing that when it was split again they would have a good shot at Wentworth.
In regards to opposition on ns it’s politically motivated on tinks part she’s effectively throwing Boyle’s chances under the bus by abolishing bradfield if the situation were reversed and boele had won bradfield and tink lost the teals would be embraced as abolition
Teals will not stand against each other.
Tell that to tink. If there was a warringah-wentworth merger I bet they would
confirmation of the aec releasing final boundaries for NSW thursday
https://x.com/AusElectoralCom/status/1832972212451881220
@John
Nice, thanks. Ready to be disappointed again, but it will be good to see the final changes.
@Angas its up North Sydney is gone
Angas and John, it looks like they only made some minor tinkering around the edges. The rest of the ‘mess’ including the awkward split of Kingsford Smith with its appendage south of Botany Bay, and the inclusion of Campbelltown council suburbs with Sutherland Shire in Hughes remains in place.
I guess these boundaries will be in place for the full seven years, unless population changes require NSW to lose another seat before then or Parliament expands in size (hopefully the latter case as that would be easier to rectify these messy boundaries by making more wholesale changes)
@yoh an i seriously doubt that it will change in less the 2 years and i doubt the parliament will be increased in the next term with the chaos of the likely minority govt and probably would be in place until after the 2028 at least and given the arrangements of the current redistributio it would be easier just to expand the parliament when they are all due for redistribution in 7 years. i think only qld and maybe wa will see a seat change trigger before then
Agree, also I don’t like how the AEC phrased the reasons why they couldn’t support the wholesale changes to Kingsford Smith, arguing that it would have flow on impacts to all divisions up to Liverpool. I would argue the whole point of this redistribution is to fix up all the districts that lack cohesion (so moving more electors to different divisions is the whole purpose, and it is self-defeating for them to do the bare minimum to achieve numerical equality).
Most redistributions have ignored the community of int
Community of interest criteria and concentrated on the numerical requirements. This is why Cook took in San souci and Hume and Whitlam.
It is interesting that Banks has moved more to earlier boundaries including part of Punchbowl. More More there is a justification to increase the number of seats in parliament this by default would increase the community of interest component.
My understanding was that community of interest was supposed to take precedence over moving as few electors as possible, according to the legislative requirements. The AEC have gone against the requirements in both NSW and VIC if that is the case (not as familiar with WA community so can’t comment on that).
Underwhelming result overall. They made some poor strategic decisions in the initial proposal which constrained their ability to make improvements at this step. Hughes and McMahon are ultimately worse than before. North Sydney always seemed like it was going to be abolished, but I thought they could do a better job about the splits of Chatswood, St Leonards and Epping, as well as areas like Blacktown and North Rocks.
Small wins:
– Amending the boundary between Dobell and Robertson so that Central Mangrove and Kulnura are fully in Robertson, to match the state district of Gosford
– Smoothing the boundary between Blaxland and Watson by swapping Watson’s share of Lidcombe and Rookwood for more of Yagoona
@Yoh An
I agree about Kingsford Smith. This was a section that they had ample opportunity to improve, but were clearly unwilling to do so.
@John @Mick
On current trends, both NSW and Vic are about 5 years away from a change in division entitlement, so these boundaries might stand for 2 or 3 elections. Maybe minority government will help to get this reform over the line sooner. I think it would be a big improvement if so.
@Adam
the written rules are Beyond the strict legislative quota requirements, the Committee’s can consider –
community interests within the proposed electoral division, including economic, social and regional interests
means of communication and travel within the proposed electoral division
physical features and area of the proposed electoral division, and
existing boundaries of divisions in the state or territory.
minimizing the number of electors move is a unwritten fourth rule
@Adam
Agreed.
My read is that the commission tends to interpret the ‘community of interest’ requirement at a ‘micro’ level in that areas being transferred only have to have a connection to part of their new division. On the other hand, I think the logical interpretation is that divisions should be cohesive at a whole, where possible.
Unfortunately, we’ve ended up with divisions with avoidable and bizarre pairings such as:
– Blacktown and Fairfield
– Coogee and Monterey
– Bundeena and Ingleburn
@angas agreed the aec did some really bad judgements this time around i reckon theyve tried to apply the so called “fairness” test and it think thats what caused these horrible divisions in all 3 states.
VIC and NSW with the obvious and WA they created a hybrid division in Bullwinkel instead of just putting it in Perths east transferring the remainder of swan from Durack and leaving Forrest and O’Connor unchanged
The solution to all this is to
NOT let the single member constituencies determine the final electorate result’s.Have an adjustment system where by there
is a extra pool of seats to balance the aberrations of
So if say Labor polls 50.8% of votes but only gets 47% of the seats then this extra pool tops up their entitlement. There would need to a threshold percentage for a party to share in this pool .
My understanding is 1940 1954 1961 1969 1980 1987 1998 elections caused this mismatch..All bar 1987 disadvantaged Labor
Other changes I think are needed are fixed term elections and deem a vote to be formal to the extent a
Voter’s intention can be identified. The later would reduce the number of informal votes.
The electorate of Bullwinkle is quite similar to the old Pearce.
@mick we live in a system of directly elected single member seats. not by percentage of vote.
otherwise basded on current polling we would be stuck with minority governemts almost indefinitely since no party has got 50% of the vote in ages.
Is majority govt worth the
Debasement of democracy? I see that this on average stopped alp govts on numerous occasions….. just think no conscription for Vietnam. No Gst…
@mick just think we could have a communist system if we just let them march across the earth
and instead of a gst we would have 100s of individual taxes instead of one sales tax
@john I really don’t think the conscription of Australian soldiers into a failed forever war saved the world from communism…
@Laine you say that now but if we just let them take vietnam how long before they march across the indian ocean. so by that argument sending troops to fight the nazis in africa and europe was a waste of time?
John, I would agree more with Laine. As a smaller country, our military forces are quite small in nature so our contribution to global battles was more of just ‘moral’ support or additional reinforcements to prop up the majority allied forces. Even for major worldwide conflicts today (Israel-Palestine and Ukraine-Russia), our military forces are just limited to providing firepower/weapons to support those actually on the ground.
Also, the Vietnam War was more of a localised conflict, similar to the events ongoing in Venezuela where it is an internal takeover by a communist group. I believe the communist forces in Vietnam weren’t keen on pursuing worldwide revolution, but instead wanted to build up localised support (similar to Venezuela and Nicolas Maduro) to rally against the West.
@yoh an they were being back and supplied by both communist china and the soviet union. before the end of ww2 the only commuist country was the USSR after ww2 it spread throughout europe, south america and south east asia. maybe you forgot that before vietnam there was korea. it wsant a localised communist push it was like a virus spreading from one country to the next we sent them there to stop its spread.
google “domino theory”
I was quite careful in my use of Language but the problem was the south Vietnamese govt was corrupt. never the less I am no fan of the communist govt of Vietnam. There are many cases of communist leaders who were unfit to lead anything let alone a country. But just because they were wrong does not make our involvement right. We did not need to be sending Australian troops there. Most wars are a tragedy and maybe with the exception of ww2 not justified.
If one believes that it ought to be an aim of an electoral system to artificially produce majority governments in defiance of how people vote, then one does not really believe in parliamentary democracy.
Sorry Laine, it did save SE Asia. The Vietnam war was a tactical defeat at the hands of Vietnamese communism, but it was a strategic victory.
You need to remember Ho Chi Minh was an old guard communist. In 1920, he was one of the founders of the French Communist Party. He fled to Moscow in 1923, where he trained at the Communist University of the Toilers of the East. He continued his revolutionary career in China, only returning in Vietnam in 1941, which he had left 30 years before. His whole life was dedicated to turning Vietnam into an independent, communist nation. He was a revolutionary to his marrow and he was also active in globalising communism beyond Vietnam.
By “bogging down” the communists in a war in Vietnam it actually saved Philippines, Thailand, Indonesia and Malaysia from either a long term communist guerrilla war (a la Colombia and FARC) or communism itself. Hence the strategic victory.
As for expansion, Laos and Cambodia fell pretty quickly in 1975….and Vietnam would itself go on to fight a guerrilla against Pol Pot until 1991.
BTW Yoh An, Venezuela since the election of Chavez’s election in 1998 has actively supported FARC and ELN in neighbouring Colombia has been actively threatening an invasion of neighbouring Guyana’s oil rich Essequibo region. It’s not contained.
Now back to NSW redistribution,
@nicolas atm the constitution sets our system as single member electorates not MMP by way of FPV so until someone changes the constitution that will be the way it goes
@john “The Commonwealth Electoral Act provides that each State and the Australian Capital Territory, and the Northern Territory on becoming entitled to more than one Member, shall be distributed into electoral divisions equal in number to the number of Members of the House of Representatives to be chosen for the State or Territory, and one Member of the House of Representatives shall be chosen for each division.46 These divisions are known as single-member constituencies. Multi-member constituencies, although allowed for in the Constitution, have not been used.47”
– via aph.gov.au
Well im pretty sure its unconstitutional because thats why those 2 hippos in Higgins were told they couldn’t job share
@Darth Vader well yeah, the country is currently divided into single-member electorates. But PR/multi-member districts are not unconstitutional (although some forms probably are due to the requirement of “directly electing” MPs).
Interesting post I remembered Raue putting up a couple years back – https://www.tallyroom.com.au/47877
The point I was making was about intention. The aim of producing artificial majority governments is counter to parliamentary democratic principles. Whatever arguments can be made in favour of single-member electorates, “it is more likely to produce majority governments” is one that ought to be dismissed.
@Darth Vader hippos? Do you mean hippies?
Last time I checked they weren’t hippopotamuses.
If you honour voter intention then you maximise the number of formal votes
The *legislation* currently requires single-member electorates. The *constitution* doesn’t require it.
It would be constitutional to have multi-member electorates using the Senate voting system as long as you didn’t cross state borders. The one other thing that wouldn’t be ideal is that the Constitution requires by-elections for the House, which would be annoying for multi-member electorates, and is different to the Senate.
Having said that, the proposed job-sharers in Higgins would’ve had a problem even under a multi-member system since their intention was to share a single seat, a single vote in parliament. Even under PR systems that elect multiple people, they each have their own individual parliamentary mandate.