We rely on independent mapmakers to draw electoral boundaries at each redistribution. There is an extensive process of public consultation, and there are rules about how seats can be drawn, but ultimately the mapmakers have quite a lot of discretion about how they apply guidelines, and where they draw the lines.
This post is looking at how boundaries are drawn federally. I have noticed some evidence of the federal mapmakers choosing boundaries that minimised the number of voters moved between electorates during the 2022-25 round of redistributions, which has sometimes produced quite strange boundaries. I think it raises some questions about how they apply the criteria, and why.
First I should clarify who I mean when I talk about “mapmakers”. For federal redistributions in Australia, there are six people who play a role in actually deciding the boundaries. The Redistribution Committee draws up the draft boundaries. This committee involves two senior AEC officials and two state public servants: the Commissioner, the Australian Electoral Officer for that state, and the state’s Auditor-General and Surveyor-General. For the final boundaries, the decision is made by the Augmented Electoral Commission, which consists of the first four members as well as the two remaining members of the Electoral Commission: the chair, who is a retired judge, and the Australian Statistician. That’s a long list of experienced and senior public servants from a variety of organisations.
The 2022-25 redistribution cycle was the biggest redistribution cycle since the first modern redistribution process in 1984. Redistributions are usually much more significant when a state or territory has a change to their seat entitlement. I call these “major redistributions”. In 2023, three states had a change in their entitlement, and those three states made up 100 out of 150 seats in the 2025 House. There was also a minor redistribution in the Northern Territory.
!function(){"use strict";window.addEventListener("message",function(a){if(void 0!==a.data["datawrapper-height"]){var e=document.querySelectorAll("iframe");for(var t in a.data["datawrapper-height"])for(var r,i=0;r=e[i];i++)if(r.contentWindow===a.source){var d=a.data["datawrapper-height"][t]+"px";r.style.height=d}}})}();
The last round of redistributions took a very long time. This next chart compares how long each stage of major redistributions have taken in the last twenty years – and also compares it to the 1984 redistribution.
The 2023-24 redistributions were some of the longest processes on record. It’s worth noting that there was a problem with the population projections for Victoria and Western Australia being incorrect, and requiring more time to use fresh projection data.
!function(){"use strict";window.addEventListener("message",function(a){if(void 0!==a.data["datawrapper-height"]){var e=document.querySelectorAll("iframe");for(var t in a.data["datawrapper-height"])for(var r,i=0;r=e[i];i++)if(r.contentWindow===a.source){var d=a.data["datawrapper-height"][t]+"px";r.style.height=d}}})}();
So there was pressure on the mapmakers to draw boundaries for over two thirds of Australia, with the schedule blowing out. That is worth noting for what we talk about next.
Prior to the last election, I calculated data on the number of voters moved from each old seat to each new seat at every federal redistribution since 1994. When I did this, I noticed that the proportion of voters moved in the last cycle was quite low compared to past major redistributions.
There have been ten cycles with a major redistribution since 1994 (there were no major redistributions in 2013). Of those ten, the proportion of voters moved ranged from 11.5% in 2022-25, up to 21.2% in 2016-19. The average was 16.2%, and no other redistribution involved less than 13% of voters moving.
This then reminded me of a number of examples of strange boundaries drawn during the 2022-25 cycle that seem to be best explained as a method to minimise voter movements.
The most blatant example is Kingsford Smith.
New South Wales was losing one seat, so most seats were under quota. This was particularly try in the peninsula covering the eastern suburbs, inner city and inner west of Sydney. The four seats of Grayndler, Kingsford Smith, Sydney and Wentworth were collectively projected to have just 3.52 quotas of population by 2028.
There was an obvious path forward – Kingsford Smith and Wentworth could both expand into Sydney, but this would push Sydney a long way into Grayndler, and also significantly shift Grayndler.
Instead, Kingsford Smith jumped across Sydney Airport to take in the bayside suburbs of the former Rockdale council area. Despite these areas being in the same local government area as existing parts of Kingsford Smith (thanks to a very controversial council merger), their connection is very weak. Unless you plan to traverse the tarmac of Sydney Airport or boat across Botany Bay, you would need to leave the electorate for quite some time to move from the old parts of Kingsford Smith to the new parts.
The advantage of this boundary-drawing is it substantially reduces the pressure in terms of the amount of population shifted through the inner west. Wentworth can actually take in parts of Kingsford Smith, rather than Kingsford Smith taking in voters from Sydney or Wentworth. Sydney still moves substantially west, but much less than it would have without Kingsford Smith’s southern push.
I actually made an objection to the NSW redistribution focused on this area. I don’t normally make my own submissions, but the Kingsford Smith change was outrageous enough that I decided to jump in. You can read it here. You can see how much more Sydney and Grayndler would have needed to shift to accomodate that Kingsford Smith change.
You could also argue that the changes to Hughes, where it shifted into the Campbelltown area, is another example of a strange boundary motivated by minimising further changes.
It’s worth noting that the mapmakers left the seat of Fowler almost entirely unchanged, despite the seat sitting in the middle of Western Sydney without particularly strong borders, surrounded by seats that required significant redrawing. It takes quite an effort to not change that seat.
Further north, some others have suggested that McMahon’s strange borders are another example – if Fowler and Reid weren’t preserved intact (seemingly because they were not under-quota), it would have been possible to draw McMahon as a far more rational seat. Instead, the seat pushed north into Blacktown, to the point that the suburb of Blacktown is divided between three seats.
Looking to Melbourne, there are some strange boundaries that seem to revolve around a strategy of abolishing the most under-quota seat (Higgins) and changing some neighbouring seats as little as possible.
There are two egregious examples in this area.
Firstly, shifting Melbourne south is understandably, but where does it go? It would have made far more sense for Melbourne to take in Southbank and other areas in the north-western corner of Macnamara. But taking in South Yarra (an area with weak connections to the rest of the seat) allowed Macnamara to be left almost entirely untouched despite having a seat abolished right next door.
The Kooyong changes are also strange. Kooyong and Toorak may be demographically similar to Hawthorn and Kew, but they are quite a distinct area. But it would have involved more voters moving to create a more natural set of boundaries.
So what is going on here? I don’t have all the answers, but I think questions should be asked of the AEC.
Is this a resourcing issue? There is a cost in time, effort and money in reorganising the electoral roll following a redistribution, and it is likely greater when more voters are moved. It is also necessary to communicate with voters when this happens.
I don’t see a particularly strong amount of political pressure from politicians to minimise voter movements – some MPs may prefer minimal change, but others would prefer greater changes.
It does appear that voter movements are reduced when the scale of the redistribution cycle is greater.
!function(){"use strict";window.addEventListener("message",function(a){if(void 0!==a.data["datawrapper-height"]){var e=document.querySelectorAll("iframe");for(var t in a.data["datawrapper-height"])for(var r,i=0;r=e[i];i++)if(r.contentWindow===a.source){var d=a.data["datawrapper-height"][t]+"px";r.style.height=d}}})}();
There is not a very strong correlation, but the redistributions that have seen greater voter movements, such as 1996, 2004, 2016 and 2019, have been when a smaller part of the country is being redrawn.
I am concerned that there may have been a shift in priorities – the two most recent cycles have seen the most conservative mapmaking in at least thirty years.
The Joint Standing Committee on Electoral Matters is currently taking submissions into their inquiry into the conduct of the 2025 federal election. I will be raising this issue, and I suggest questions should be asked of the AEC. Are there resourcing constraints forcing the drawing of less logical boundaries? Or is it a time constraint, worsened by the ABS projection problems seen in the last cycle.
Right now it looks like the 2025-28 cycle will be a relatively modest one for redistributions, but this issue will arise again in the future. And if the parliament is expanded, it’s important that the mapmakers have the capacity to draw the best boundaries, and not just play it safe.
I asked the AEC media team for comment on this issue but they declined to comment.