The Australian Bureau of Statistics yesterday released its latest quarterly population estimates (as of last December) for Australia and its several states and territories. The country’s population was a little over 25.52 million, representing an annual growth rate of 1.4%.
This would be of no more than passing interest were it not for the fact that, because it is twelve months since the current federal parliament sat for the first time, these are the statistics that will be used to determine each state and territory’s representation in the next parliament. Where that entitlement has changed, a redistribution will have to be conducted.
As we noted last year, while Australia’s electoral boundaries are drawn on the basis of equalising numbers of enrolled voters, the apportionment of seats to the states is done by total population – as set down in section 24 of the constitution. And in 1975 the High Court ruled (in McKinlay’s case) that it wasn’t enough to just do this after every census; it had to be done for each parliament (that is, at least every three years).
So in two weeks time, on the basis of these figures, the electoral commissioner will determine each state’s entitlement. Victoria, with high population growth, will gain a seat, going to 39; Western Australia, with relatively low growth, will drop from 16 back to 15. The other four will be unchanged. (Tasmania by population would only be entitled to three seats, but as an original state it is guaranteed a minimum of five.)
The constitution doesn’t prescribe how the territories should be represented; that’s a matter for parliament to determine (section 122). Parliament has legislated to apply the same formula as for the states, and on that basis these figures would have the ACT lose one of its three seats and the Northern Territory lose one of its two.
But there’s a catch, introduced in 2003 when the Northern Territory was at risk of losing a seat. In dividing by the quota, an allowance is made for statistical error in the population estimates. That was enough to save the third ACT seat last time around, and it will do so again.
The Northern Territory, however, is well short even with that allowance: its population is only 1.433 times the quota. Unless parliament legislates to make some new provision, it will drop down to just one seat.
The Territorians, of course, are not at all happy with that idea. They argue, with some justice, that it is grossly unfair that they should get only one MP for almost a quarter of a million people while Tasmania’s five will represent only about 107,000 each.*
In a recent series of posts, Antony Green argues for a simple remedy. He suggests that the proportionality should be calculated by using the harmonic mean rather than the arithmetic mean – that is, by rounding a state or territory’s entitlement in the direction that will minimise the difference between the resulting average population per seat and the national average.
For the states, it would almost never make a difference, since the two methods converge for larger numbers. But for the small numbers involved with the territories, it would mean a significant boost, rounding at 1.33 and 2.4 rather than 1.5 and 2.5.
I think this is a very sensible idea, and I hope parliament listens to Green. If you’re interested, his posts contain a wealth of extra detail about how the different methods work and how it relates to the similar debate about Congressional apportionment in the United States.
The underlying problem, however, doesn’t go away. Because the territories are so small, there would still be a point at which a small difference in population would make a very large difference to their representation, and one day the same situation will recur. Tasmania’s guaranteed minimum entitlement just makes the injustice more apparent.
My solution would be more radical. The constitution mandates (section 29) that no seat may “be formed out of parts of different States,” but there is no such restriction on the territories. Therefore, it would seem, there is no constitutional barrier to merging the territories for the purpose of representation – either with each other, or with a state.
So, for example, if we simply add the ACT to New South Wales and the Northern Territory to South Australia, neither territory will be liable to any sudden sharp change in its representation. On yesterday’s figures, the former combination would be entitled to 50 seats and the latter to 12, in each case requiring a seat to cross the state/territory border.
While I think this would be a major improvement on the current system, I have no illusions that it’s likely to be adopted. The big thing for territories is to try to convince everyone that they are as much like states as possible, so they are unlikely to ever sign up to a proposal that might make them look like just an appendage to a state – even if it would sometimes work in their favor.
* That said, if you look at enrolled voters rather than population the picture changes a bit, and in particular the relativity between the ACT and the Northern Territory is quite different: at the last election the ACT had 98,600 voters per seat, while the Northern Territory had only 69,700.
4 thoughts on “Representation games in Australia”
Drawing Queanbeyan into an ACT division makes enormous sense. But the great outback electorate centred on Alice Springs would probably rival the old Kalgoorlie for size.
Yes, the ACT would be pretty straightforward. The outback seat would be a sort of Lingiari/Grey hybrid; it’d be big, but it’d have a coherence that both of them lack, not having to accommodate either the Darwin outskirts or the mid-north of SA.