In yesterday’s post I explained my dataset of election results from 2004-2019 adjusted for 2022 boundaries. In today’s post I use that data to explore how seats have changed their political position relative to the rest of the country over this period.
Every electorate has a two-party-preferred (2PP) count, which is what happens when you distribute all preferences to either the Labor or Coalition candidate. The AEC even conducts a 2PP count in a non-classic seat, where the two-candidate-preferred count involves an independent or a minor party.
This means that you can rank all 151 electorates according to their relative 2PP. For this exercise, I’ve ranked the seat with the highest Labor 2PP as 1, and the highest Coalition 2PP as 151. There will always be a median electorate which is ranked 76th. The seats under 76 have a relative skew to Labor, with those over 76 skewing to the Coalition. Of course this doesn’t mean Labor won those seats under 76: if the Coalition wins a big victory, the 75th best Labor seat would probably be won by the Coalition. And of course some seats are won by crossbenchers.
When there is a national swing one way or the other, it’s quite common that most seats will swing in the same direction, but swings are never uniform. Comparing the percentage vote requires some understanding of the national baseline to make a fair comparison of how much each seat moves, but a ranking is simpler, and doesn’t require that comparison.
First up, this map shows each seat’s rank at each of the six elections in the dataset.
Feel free to wander around on that map and look at seats and regions you find interesting, and posts your thoughts in the comments.
But to summarise the trends, I wanted to look at the seats with the biggest changes in their rankings. This next map shows this in some way. The first layer shows how much the rank has changed from 2004 to 2019, with seats shifting towards Labor in red and the reverse in blue (those not moving much are yellow).
The second layer shows the largest net change over that period. Some seats peaked in years other than 2004 and 2019. For example, Capricornia has shifted from 51st for Labor in 2004 to 118th in 2019. It was the median seat in 2016, before finally shifting clearly into the Coalition-leaning seat list in 2019, but it was most pro-Labor in 2007, when it was ranked 33rd.
The seat of Capricornia has shifted furthest to the Coalition between 2004 and 2019, shifting 67 places. On the other hand, Mayo has shifted 51 places towards Labor. It was Labor’s 127th best seat in 2004, and it was the median seat in 2019, a shift of 51 places.
Other seats shifting 40 or more places towards the Coalition were Bonner, Gippsland and Calare. Kingston has shifted 47 places towards Labor, and Warringah has shifted 40 places towards Labor. I was amazed to realise that it was Labor’s 75th-best seat on two-party-preferred vote in 2019, shifting from a 115th-best in 2004 and 134th-best in 2007.
If you just look at the biggest shift, rather than just the biggest 2004-2019 shift, Capricornia has moved 85 places. Wentworth has moved 63 places, from being Labor’s 81st-best seat in 2004 after Malcolm Turnbull’s controversial preselection challenge to his predecessor. It eventually was ranked 144th in 2016, before becoming less pro-Coalition in 2019. Dawson has also shifted a long way. It was Labor’s 75th-best seat in 2007, but was 134th in 2019.
Finally, I wanted to look at the seats that have shifted between the Labor and Coalition sides. I looked at seats that shifted from Labor’s top 75 in 2004 to the Coalition’s top 75 in 2019, and vice versa. Nine seats have shifted in each direction.
Banks, Bonner, Braddon, Brisbane, Calare, Capricornia, Lindsay, Reid and Swan have all moved out of Labor’s top 75. Some like Lindsay, Braddon and Swan have only moved a small distance towards the Coalition, but others have moved by at least 30 places.
Boothby, Dobell, Dunkley, Eden-Monaro, Gilmore, Macquarie, McEwen and Warringah have all moved into Labor’s top 75. Corangamite was the 76th seat in 2004, but now leans further to Labor, while Mayo has moved from being on the Coalition side to the 76th seat. More of these seats have only shifted a short distance, but Dunkley, Gilmore, Macquarie, McEwen and Warringah have moved by at least 30 ranks.
When you look at electoral trends over longer time periods, the familiar electoral battlegrounds tend to change. Unfortunately I only have the matching data for the last 15 years, which is not quite long enough to detect long-term trends, but when you read the histories of each seat you see seats that were once key marginals, but are now much safer. I think we can see a small taste of those trends in the 15 years worth of data, and I look forward to expanding it in the future.
As a last item, you can also click through the following map which is coloured by my estimate of the 2PP in each 2022 seat at each of the six elections.