Enrolment growth forces NT redistribution back to the drawing board


The Northern Territory Electoral Commission published its draft boundaries for the 2024 Legislative Assembly election back in May, but that process will now have to start over due to tremendous enrolment growth in some parts of the Territory in recent months in the lead up to the Voice referendum.

The NTEC announced two weeks ago that there had been a substantial amount of enrolment growth concentrated in some parts of the Territory from mid-April 2023 (when the quota for the redistribution was struck) until the end of June 2023.

Overall the size of the electoral roll grew by 4,631 voters, which was 3.1% growth across the Territory. The average electorate has an enrolment of roughly 6,000 voters, so that is quite a substantial increase.

What makes this particularly difficult is that the enrolment growth is concentrated in some areas, although it is a good news story as it suggests an improvement in the dismal enrolment rate in the federal seat of Lingiari.

The growth in urban areas (Alice Springs, Darwin and Palmerston) was quite modest, at just 1.4%, while the remaining nine seats grew by 6.3%.

Indeed if you exclude Goyder and Katherine, which experienced only modest growth, the remaining seven seats increased their enrolment by 7.7%.

68% of the growth during those two and a half months has been concentrated in those seven seats. Those seven seats have gone from 27.8% of the territory total enrolment to 29% in just ten weeks.

This growth has apparently been caused by a major push in the Federal Direct Enrolment Update (FDEU) program from the AEC. We don’t yet have the AEC’s data on enrolment by division in June, so I haven’t been able to see if there are broader trends. Up to the end of May, Lingiari appeared to be one of the few seats to have a shrinking roll.

At an individual level, this will definitely cause some of the newly drawn seats to no longer fit within the quota. The NTEC has published the updated enrolment statistics based on the old 2020 boundaries, not their previous draft 2024 boundaries, but we know that the seat of Mulka, which was previously 9.2% above the average and was left untouched by the redistribution, is now 15% above the average and will definitely require a redraw.

Individual electorates can be above or below quota without causing broader issues, if that population deviation can be balanced out by the neighbouring seat. But when whole regions are collectively under or over quota that forces bigger changes.

For this next chart, I’ve grouped electorates in the same categories I used when analysing the older enrolment data in March.

Region Seats April deviation June deviation Enrolment growth
Alice Springs 2 0.4 -0.8 2.5
Darwin 9 -19.0 -37.1 1.0
Palmerston 5 22.4 14.3 1.5
Urban 16 3.8 -23.7 1.4
Outback 3 -14.7 2.5 9.3
Top End 6 10.9 21.2 4.9
Rural 9 -3.8 23.7 6.3

On the previous figures, urban NT was collectively about right for size while rural NT was also about right. Darwin was significantly underpopulated but that deviation was pretty much balanced out by Palmerston. The retraction of Spillett into being a Palmerston-only seat basically covered that issue without causing larger knock-on effects.

That is no longer the case. Palmerston is still over quota, but less so, while Darwin’s shortfall has doubled. Meanwhile the southern outback seats were previously under quota but have been brought up to quota, while the surplus population in the six top end seats has doubled.

Overall this means there is almost a quarter of an electorate’s surplus enrolment in those nine seats. If you narrow the scope to just Gwoja, Barkly, Mulka, Namatjira, Arnhem, Arafura and Daly, they are collectively over quota by 25.1% of a seat.

I won’t try and speculate about particular population changes but ultimately this means those rural seats need to shrink, shedding excess population. In particular Mulka needs to lose a lot of its voters to its only neighbours Arnhem and Arafura, which will likely cause knock-on effects into Goyder.

Palmerston will still be able to partially deal with Darwin’s deficit by having a seat like Spillett shrink, but less than before. Ultimately those Darwin seats will need to expand in an eastward direction to absorb that surplus enrolment from rural areas.

There’s also a question about whether this process has concluded or whether this change in enrolment trends could extend further.

The AEC estimates that just 89.4% of potential electors in the Northern Territory are enrolled to vote. This is much lower than any other jurisdiction, but has improved remarkably in recent years. The enrolment rate in the NT was under 80% as recently as March 2016.

The AEC doesn’t publish precise enrolment rates per division, and the NTEC certainly doesn’t do so for the smaller NT Assembly electorates, no doubt due to the imprecision in estimating resident populations between censuses. It’s much easier to count how many voters are actually enrolled to vote. But the AEC as of May estimated that 80-85% of eligible electors in Lingiari are enrolled, compared to over 98% in Solomon.

Just three seats in Australia have an enrolment rate under 90% – Lingiari is joined by the massive WA seat of Durack in the 80-85% bracket while Melbourne is at 85-90%.

Overall this suggests there is still plenty of potential for enrolment growth in the NT to be heavily slanted towards the remote electorates. If that growth continues, it could force further redistribution changes that shift power away from Darwin. In just ten weeks, these electorates went from having no enrolment surplus to having a surplus equivalent to a quarter of a seat. I wonder how much is left to do. The NTEC will need to consider what is left in this program, and if there is more to come they might want to err on the side of drawing those remote electorates at the bottom end of the permitted range to allow for growth. Such a decision would force more dramatic changes between the Darwin area and the remote electorates.

Finally, I have made a map with three layers: the deviations by seat on the April figures, the same deviations on the June figures, and how much each seat’s enrolment has increased in the intervening ten weeks.

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


  1. The location of the enrolment suggests that indigenous people are making a big push to get on the roll for the Voice referendum. Post 1 July there will be 500,000 kiwis getting on the roll to vote yes also?

  2. You mentioned that enrolment rates are very low in three electorates, two of which (Lingiari and Durack) are outback and have significant Indigenous populations. But the third is “Melbourne at 85-90%”. It is good to hear that rates of enrolment in Lingiari are improving, but what is the reason for such low rates in Melbourne? Is this a tendency by younger voters to ‘disengage’ with voting?

  3. @Stephen Morey yeah Melbourne is a weird one. What’s the enrolment rate in other remote electorates with high Indigenous populations (i.e Grey, Kennedy, Leichhardt, Maranoa, O‘Connor and Parkes)? Although Leichhardt does have Cairns, a major city, so maybe Leichhardt is higher but there is still lots of Aboriginals and Torres Strait Islanders (as well as Papuans) in Cape York and the Torres Strait Islands.

  4. The enrolment rates by Division can be found at https://www.aec.gov.au/enrolling_to_vote/enrolment_stats/rate-div/index.htm
    This is arranged by State and in Victoria all other divisions have more than 95% enrolled and several over 98%. Melbourne stands out.
    Kennedy and Leichhardt are both between 90-95%. But it’s possible the enrolment rates are much lower in the Indigenous areas of Leichhardt, I have no information on this. Ben might know.
    Indigenous enrolment by State / Territory is here: https://www.aec.gov.au/Enrolling_to_vote/Enrolment_stats/performance/indigenous-enrolment-rate.htm
    It looks as if the AEC has been doing a good job at increasing Indigenous enrolment since 2019. Indigenous enrolment rates are lower in WA than in NT.

  5. the one roblem i had with the original redistribution is the fact they refused to divide mulka even though it was over quota simply becuase they didnt want to split up language groups. thats the same as saying they dont want greek and turkish voters in the same division in melbourne. last i checked they are meant to make divisions as equal as possible not put all the same people together imagine if they let a division in victoria stay close to 10% over quota simply becuase they didnt want to divide a specfici cultral group

  6. @vicliberal Is that true? Honestly I thought Mulka was entirely on Yolŋu Country and they spoke Yolŋu and English. However there are different Yolŋu dialects. In fact the “mulka” itself is Yolŋu for “dry area”, but it is also used to describe “a diplomatic space where people are welcome to visit and bring business and discussions”, as the NTEC states on their page about Mulka (https://ntec.nt.gov.au/electoral-boundaries/legislative-assembly-divisions2/division-of-mulka). Anyway it is weirdly shaped, if you look at the map at the bottom of the page I just linked. It was known as Nhulunbuy until 2019. The CLP needs to contest this seat, it’s weird for a major party not to contest every seat in an election even if they can’t win. Interestingly in 2016 the CLP’s candidate was Charlie Yunupingu. Is he related to the other Yunupingus?

  7. The other thing they should look at is the elongation of gwoja surely something can be done to take Barkley and Namatjira and create 3 districts of better size?

  8. @Vicliberal you mentioned the ethnic issue in Mulka earlier. Well, just to emphasise the fact that dividing electorates by ethnicity instead of population is a stupid thing to do (as you pointed out), here is what the NSW state electorates of Lane Cove and Ryde in Sydney would look like if they were divided by ancestry according to the Census (i.e Australian, Chinese and English): https://www.mediafire.com/file/uagis2ijx894849/B521B163-D68D-459D-A621-6D65484DECE7.jpeg/file https://www.mediafire.com/file/jinvit4yh17hnzg/6F35E195-6630-4BB3-AB6C-0A26524AF9FB.jpeg/file

    So basically the federal electorate of Bennelong would be split into at least three electorates if it was by ancestry, plus whatever the demographics of Epping are.

Comments are closed.