Will the NT lose one of its federal seats?


I have been meaning to write a post about the impending determination and possible solutions to the likely merger of the NT’s two federal seats. Antony Green put together a long and interesting piece yesterday explaining how the current formula short-changes the territories, and suggests an alternative formula which would produce a better result, as well as some background on where our current system comes from. I’m going to quickly run through this issue for those who don’t have time to dive in deep.

I wrote about this likely outcome in August last year. It looks likely that Victoria will gain a 39th seat, Western Australia will lose its 16th seat, and that the Northern Territory will revert to just one electorate, after gaining a second seat at the 2001 election.

The current formula determines a national quota dividing the total population of the six states by twice the number of state senators (144). This formula is the default specified in the Australian constitution (although the parliament does have some room to move by changing the formula). The same formula is used to allocate ACT and NT seats, but this is not fixed in the constitution and could be changed.

Labor senator Malarndirri McCarthy has put forward legislation that would set a minimum number of electorates for the territory at two, thus averting the seat loss. But Antony has proposed another change which would slightly improve representation for the territories without setting a hard floor.

This formula can be problematic when a territory has a small number of electorates. There is a massive change in the average population per seat when you jump from one to two seat, or from two to three. The average population per seat will be much higher if you have 1.4 or 2.4 quotas of population than if you have, say, 40.4 quotas.

Antony recommends a formula called Dean’s formula.

In short, it allocates the number of seats to each jurisdiction which would bring the average population per seat as close as possible to the national quota.

In the case of the NT, this means that anything more than 1.33 quotas of population would qualify for two seats. A third seat would be allocated if a territory had more than 2.4 quotas. This formula gets pretty close to the current formula once a jurisdiction is allocated more than a few seats. It’s actually simpler than I thought it would be, and I think it could be a good solution.

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


  1. Either we live by the principle “One Vote, One Value” or we don’t.
    Under the current guidelines, the NT will inevitably revert to a single Division at the Determination due next month. And, as someone who’s been doing election analysis & contributing to electoral redistribution’s at both Commonwealth & State/Territory levels for the past 7+ years that’s fine with me. The turnout at the 2 NT Divisions is usually the lowest in the country.
    Let’s not go down the path of making special laws for the territories compared to the States.
    It’s bad enough that Tasmania has to have 5 representatives due to Section 24 of the Constitution when it’s only numerically entitled to 3.
    However, there is a way to ensure that the NT remain entitled to keep their current 2 Divisions – and it’s been done before.
    Simply increase the number of Senators from the States from 12 to 14. Problem solved. Yes, it will force Redistributions for all other States except Tasmania but Victoria and WA were going to have them anyway, and NSW is due to commence in 2022 – but would have to wait until after the next election.
    The last time we increased the size of the parliament was under Hawke in 1984 which was 35 years after the first expansion under Chifley in 1949. It’s been 36 years since the Hawke expansion so, given how much our population has grown since 1984, I think we’re due again.

  2. There is no need to make a special formula. If you use Dean’s formula, which uses the harmonic rather then arithmetic mean as the round off point on fractions, then it can also be applied to the states. I’ve been through 26 state apportionments and it changed one state’s allocation by one seat at one apportionment. The harmonic mean has the property that it minimises the differences between average population per member in states and territories. It can be easily argued it is more proportional than the “until parliament otherwise decides” formula in Section 24 and therefore the High Court would allow if legislated.

  3. Seems out of line for such a large part of the country to only have one HOR representative no matter the turn out. So whichever method keeps their representation up, by simplest formula, is fine by me.

  4. Jeff Waddell writes: ‘Simply increase the number of Senators from the States from 12 to 14. Problem solved.’
    This would worsen the over-representation of the small states. For example, currently the ACT has 80% of the population of Tasmania but only 5 federal representatives compared to Tasmania’s 17. Let’s not increase that to 19.

  5. Really looking forward to Antony’s Part 3 and any thoughts he has on whether the Dean formula could be applied to the states without running into the issues identified in the McKellar case. The concern would be that because it can yield a threshold for the extra seat that is below 0.5, but never above, it “biases high” on the number of seats and is more likely to yield seats for the states totalling more than 144 (before Tasmania is bumped up to its minimum five) than it is to yield less than 144. So it might not pass the “as nearly as practical” test. But unlike the “always round up” provision considered in McKellar, the Dean method might be argued to improve on the “in proportion” part of the provision. So maybe it could get through.

    My understanding of Huntington-Hill is a little different to Antony’s description, but his description might be specific to the territories, with no cap on the number of seats to be allocated. I think Huntington-Hill as used in the USA for the House of Reps (allocating a fixed number of seats) would definitely pass the McKellar test, but the concern would be whether it would be volatile and cause more frequent redistributions because it would insist that if one state gains a seat, another state must lose one

  6. Huntington-Hill passes the McKellar test because at all 26 apportionments since Federation it produces the same result. Dean produces a one seat difference in one state at one apportionment. Both methods have a significant impact on the territories though.

    But because of the way the Dean method works, that one seat difference in one state produces an average enrolment for the state that is closer to the national quota than does the Section 24 method. For that reason I would argue the Dean method more proportional than Section 24, and as Justice Stephen said in McKellar, proportionality is absolute terms and the nexus is in qualified terms.

  7. Section 24 has no cap on the number of seats to be allocated. All of the literature on the US methods assumes a cap, and also assumes a minimum of one seat. The allocation tables used by Huntington-Hill are not necessary if you don’t have a fixed number of seats so you can just round all the seats at the geometric mean.

    Websters method in the US is also known as Sainte Lague and works off allocation tables. We use Webster’s method in Section 24 without the allocation tables because there is no fixed cap on seats.

  8. I like the Dean method and think it optimises what we should be optimising. I’m just less sure that it is a perfect fit with what the language of section 24 requires us to optimise, i.e. that members be proportional to population. So, as much as I think a state (or territory) with 1.4 quotas should get 2 members because 0.7 is closer to 1 than 1.4 is, I’m struggling to make the case that 2 is more proportional to 1.4 than 1 is. I hope I’m wrong!

  9. Boo Hoo NT loses a seat . i have a much better solution ABOLISH BOTH TERRITORIES. And SA, & TAS while we are at it. The NT also has 2 YES TWO senators totalling 4 YES FOUR elected representatives for about 150000 voters FFS !!. Not good enough. TOTALLY RIDICULOUS.

  10. Dean, it is the average number of people per member in a state or territory that is closer, not the quota size. In NT, the average with 1 member based on 2017 numbers was 247,512, for two members 123,756. The National quota was 164,789. At 1.33 the average enrolment produced by division is 185,634, the mid-point between the average number at 1 and 2. Below 1.333 that divisor is closer to the one seat value, above 1.33 it is closer to two seats.

  11. Dean’s method minimises the difference between states in average people per member. Webster’s method minimises the difference in members for population, and Huntington-Hill doesn’t use absolute proportionality, it looks at relative proportionality.

  12. Antony, my point is that section 24 says members should be proportional to population, not vice versa, and definitely no mentions of averages. As such, while I much prefer the outcomes from Dean, I can see an argument that Webster is the closer fit to the wording. Like I say, I hope I’m wrong.

  13. You are saying that if ‘a’ is the number of seats in a state, and ‘p’ is its population, then any method you adopt must minimises the differences in ‘a/p’ across the states. Dean’s method minimises differences in the inverse ratio, ‘p/a’, population divided by members. Webster’s method as used in the USA, and as written in Section 24, is the only method that can minimse ‘a/p’, so by your logic the formula cannot be changed. But there are two problems with that. The Constitution’s writers did say until parliament otherwise decides, and why would they say that if as you suggest the wording says you can’t use a method other than one minimising ‘a/p’. So are you constructing a meaning that wasn’t meant. Second, the a/p minimisation works where you have a minimum 1 per state and a fixed size House. Neither of those conditions apply in Australia. In my view you go back to arguing the meaning of proportional.

  14. The NT used to have lots of jobs for unskilled citizens be they have disappeared like in other party of Australia. The aboriginal population has increased and as a result more professionals like doctors reside in the NT. However many of these professionals are New Australians who when completing the obligation to work in remote areas then mover to the southern capitals. Even moving a large part of the ADF to the NT in the 1980’s has not helped stabilize the population much either.

  15. John
    Why ? I struggle to see any benefit. Don’t we just end up with more useless politicians ?

  16. better idea to increase the number of politicans……..too many people per electorate…. the quota is approx 100000 voters

  17. Mick, I agree.
    At the last expansion under Hawke, there was an average of 73,414 electors per Division BEFORE the expansion & 62,158 after.
    Other Stats that I think are important: At the 2019 Federal Election, there were 21 Divisions where the turnout of electors to elect a SINGLE MEMBER to the House was greater than the turnout of the entire NT to elect 2 members. How anyone can define the continuation of this arrangement as “fair” is beyond me!
    In the past 3 years – based on the quarterly population stats as published by the ABS, the population of Australia as a whole has increased 1,136,564 & the population of the NT, Cocos (Keeling) & Christmas Islands has decreased 232.
    I’ve lodged a submission to the JSCEM relating to the proposed amendment which details these stats and more.

  18. Jeff Waddell /mick quinlivan
    Yeah fine. BUT for what benefit ? What will be achieved?. It is not the cost of politicians it is their staff, & everything else. I want REAL Representation. Not more better, & different of the same. Bugger that !

  19. winediamond – I hear you. And this is a whole different issue. The problem is that the political establishment have assumed that the power that comes from us at the ballot box is actually theirs. It’s been a duopoly for 3/4 of a century and no-one else gets a look in. And when someone tries like the Democrats at the end of the last century they might survive for a while, but the establishment wears them down.
    We’ve got Labor with branch-stacking and alleged CCP influences and a Liberal-National party that’s generally weak. Both parties are guilty of not putting Australia or Australians first in their policies.
    2 years ago, when my IT job got sent off shore and my 22-year IT career with it, I wondered how both sides could allow the people that vote for them to be treated that way.
    I gave up on both of them and joined One Nation. Even stood in Monash at last year’s federal election. We might not be perfect, but we do believe in taking the power of the political class away from them and giving it back to the people. We just need enough people to realise how they have had their rights and liberties taken from them and realise that all the main parties are in it for themselves, not the people they’re supposed to represent.
    It’s a long road back from where we are, though.

  20. Jeff
    I am with you all the way. I don’t disagree with anything you have said, & could go on forever, with more.

    Well done to you for standing, & having a crack. Might i suggest Holt may be more fertile ground next time !.The member seems in some difficulty……!.

    When people accuse me of allegiance .>>”i vote AGAINST politicians, NOT for them (like any sensible person!). I voted for Latham in 2019, because he EARNED my vote. Something no one has done since Ted Mack.

  21. @Jeff Waddell. Regarding the turnout in the NT, it’s obviously a combination of low turnout (78% in NT vs 90%+ in every other state and territory), under-enrolment, and demographics (NT population is slightly younger, so more under-18s).

    I guess the question is, do MPs represent people, or just voters? I’m in the former camp. I’d actually favour changing the redistribution methodology to be based on population rather then electors. As compared to the current hybrid where the constitution requires that the allocation of divisions to states be based on population, but the Electoral Act then requires that divisions within the state be equal on enrolments.

Comments are closed.