From one election to the next, the swings back and forth of Australian politics can look cyclical – that things start and end in the same place. But when you widen the time frame, things look different. Over time, some place move from being safe seats to marginal seats, or even from being safe seats for one major party to the other major party.
Back in 2022 I published data that I produced converting federal election results since 2004 into modern electoral boundaries. I have now converted that data to the 2025 boundaries, and added the 2022 election results. That’s seven elections worth of data that is comparable in a way you can’t do with standard election results, due to the impact of redistributions. Throughout this blog post, all references to electorates will be based on the current 2025 boundaries.
With two-party-preferred votes, you can rank every seat in the country from 1 to 150, which makes it possible to compare different elections where the country as a whole may have been leaning more towards Labor or towards the Coalition. For this post I am most interested in seats that have shifted one way or the other, and particularly seats that have moved in or out of the group of seats that decide elections, primarily comparing the elections where Labor came into government: 2007 and 2022.
When I compare the 2022 results to previous elections, unsurprisingly I find the correlations weaker as I go further back in time, but actually 2007 is more unlike 2022 than 2004 is. 2004 was an election where Labor gained ground in the inner cities while losing ground elsewhere. It was the election that coined the term “doctors wives”. Indeed while reading the 2004 book Mortgage Nation I came across this table from Malcolm Mackerras. Most of the seats on this list have been lost for the Liberal Party, many of them to teals.
The 2007 election had a Labor 2PP of 52.7%, while 2022 had a Labor 2PP of 52.1%. Both elections saw the election of a first-term Labor government. In 2007, Labor won 83 seats on the 2PP. In 2022, Labor won 77 seats, plus seven crossbench seats had a Labor 2PP majority, which adds up to 84 seats. While there were major differences, the two elections are very comparable.
I will post maps below showing how much each seat shifted between 2007 and 2022, but first this chart shows the rank of every seat on the 2PP in 2007 and 2022.
Seats in red have changed their rank in Labor’s favour over these 15 years. Seats in blue have shifted towards the Coalition. Overall 71 seats have moved towards the Coalition and 79 towards Labor. The black diagonal line indicates where every seat would be if they kept their rank, so seats close to this line haven’t moved much. If you’re closer to the top left corner or bottom right corner, the seat has moved a lot.
The horizontal and vertical lines mark out the seats that ranked between 66 and 85 in each of these elections: the seats that decide the election. Amazingly, only one seat was in both lists: McEwen, in Victoria. That seat was Labor’s 82nd best seat in 2007 and is now 71st-best, on the 2025 boundaries.
The seat that has shifted the most towards the Coalition is Capricornia. This seat was a safe Labor seat in 2007 but is now a safe LNP seat. It moved 82 places over those 15 years. On the other hand, Pearce in Western Australia has moved 69 places. Capricornia has shifted gradually but most of Pearce’s change happened in one election. The seat was ranked 120th in 2007, was ranked between 86 and 99 from 2013 to 2019, but then dropped to 51st in 2022. Capricornia’s change seems more likely to be permanent, while Pearce may well fall back into the pack.
In terms of marginal seats, Kingston, Perth, Franklin and Burt were all in the 20 most marginal seats in 2007 but are now amongst Labor’s 21 safest seats. Page, Dawson and Braddon were also in that cluster of important seats but are now amongst the Coalition’s 31 safest seats.
Hunter, Fowler and Werriwa were all amongst Labor’s 22 safest seats in 2007, but are now close to or in the 20 most important seats. Mayo, Bullwinkel and Tangney were all amongst the Coalition’s 36 safest seats but are now in the key 20 seats.
When you look at seats by their regional classification, quite a lot has changed. When you split the country into Labor’s 75 best seats and the Coalition’s 75 best seats, Labor has made a net gain of 3 seats in inner metro areas, and a net gain of two in outer metro areas. The Coalition has had a net gain of one in provincial areas and four in rural areas. Over just 15 years, it is possible to identify the urban-rural trend which has defined the last century. Labor improved their position in 29 out of 43 inner metropolitan seats, while the Coalition improved their position in 22 out of 38 rural seats.
Finally, this map shows the relative shift in ranks for each seat from 2007 to 2022. If you click on a seat you can see my estimate of the 2PP for 2025 boundaries for all elections back to 2004.
Broader trends become obvious on the map. WA in general has shifted strongly towards Labor, while Queensland has mostly shifted to the LNP. The south of Tasmania is more pro-Labor while the north is more pro-Liberal. In the south-eastern states, cities have moved to the left while rural areas have moved to the right. But there are exceptions in the cities – the south-west of Sydney and the west of Melbourne are substantially more friendly to the Liberals than they were in 2007.
This story doesn’t end in 2022, and it will be interesting to see which seats will swing more strongly in 2025. The marginal seats of tomorrow will not be the same as the marginal seats of the past.
I think the penultimate paragraph has a typo: “Queensland has mostly shifted to Labor”
I would take issue with the idea that
CAPRICORNIA is now a safe lnp seat.
The margin is 6% odd. This is possible but
Unlikely to flip.
I get your point Mick, I guess the point is more that it is a long way away from being a seat that would decide the election. If Labor gets a 6.6% swing they are going to win by a lot. It would be different if the margin was 6.6% when Labor was at a low point. The margin is 18.6% weaker for Labor than in 2007.
I am loathe to transpose state results onto federal seats, but the three seats that take in Capricornia (Rockhampton, Keppel and Mirani) all saw massive swings away from Labor. Mirani dropped 11.6%, Keppel 16.7%, and Rockhampton 14.1%. Add in the fact that Landry is now entrenched as the incumbent, this seat is her’s until she retires. Even then, it may take some special alignment in the electoral calculus (another Queenslander as PM?) for Capricornia to return to the ALP. This seat is also particularly fertile ground for One Nation, with nearly one in six electors putting the orange candidate first in 2022, and a high profile campaign in Keppel in the state election.
Agree Real Talk, also the fact that Labor retained Rockhampton and Mackay during the landslide 2012 defeat (when they lost almost every Brisbane area seat) further shows that these seats have diverged and are now conservative leaning.
Conversely, some of the inner-city and middle ring districts are now more favourable for Labor. Clayfield district still remains marginal (<5%) even as the LNP won office. Cooper (formerly Ashgrove) also remains Labor held with a margin above 10%.
I agree Real Talk, this seat is Landry’s until she retires. Talking to people in Capricornia, apparently Russell Robertson (2019 & 2022 Labor candidate) had some issues and was a particularly poor choice.
From what I’ve heard, Labor retaining Mackay in 2012 was partly to do with Tim Mulherin’s strong personal vote. Similarly, Cooper in 2024 has such an inflated margin because of Jonty Bush’s strong personal vote. Plus the LNP didn’t try in Cooper at all.
@Mick
There is no chance Labor would pick up Capricornia. Michelle Landry has a personal vote there, and Matt Canavan has been campaigning heavily on local issues too. The strong One Nation vote (funnelling preferences to the LNP) makes it even harder for Labor to pick it up. Plus, Labor’s candidate lived in Brisbane until she was preselected, which isn’t a good look in the regions. Capricornia is a seat where Dutton appeals to voters. It’s a regional seat with a lot of ‘battlers’.
Something that I think will help Labor is that marginal seats seem to have good trends towards Labor. Of the 38 seats that had ALP vs L/NP margins under 5%, 23 were in the trending to Labor column, 15 to the L/NP. It gets even more drastic when you look at the ultra marginals (under 3%), of the 21: 15 are trending to ALP vs 6 for the L/NP.
If you look at marginal Labor seats under 5%, these are the one’s that are shifting Liberal
Lyons (.9%)
Lingiari (1.7%)
Paterson (2.6%)
Parramatta (3.7%)
Hunter (4.8%)
Now Liberal seats with margins under 5% that are shifting Labor
Deakin (.02%)
Bennelong (.04%)
Sturt (.5%)
Moore (.9%)
Canning (1.2%)
Casey (1.4%)
Monash (2.9%)
Forrest (4.2%)
Durack (4.7%)
+ the IND seats that they could eventually win the 2pp
Warringah (.7%)
Wentworth (1.2%)
Goldstein (3.7%)
Kooyong (3.7%)
If these trends continue, any sort of status quo result or even just a small swing to the L/NP will end up helping Labor. We saw this thing happen in Victoria where a lot of marginal seats are trending towards Labor and it’s more-so safer seats to the Libs.
Drake, it’s an interesting point you raise considering the latest MRP had a 50.2% Labor 2PP and projected 75 seats vs 60 for the Coalition. It seems that YouGov projects a favourable swing distribution to Labor in that MRP. The last MRP was the other way around though – where the swing was mapped as being more favourable to the Coalition. Small shifts in the vote on the surface may be hiding bigger underlying trends.
The MPR before the previous one was more favourable to the Coalition in swing distribution, I mean.
Mrp is I think good for global totals only.
When is next opinion poll due?
@Drake, Deakin, Bennelong and Sturt have sizeable Chinese Australian which shifted way from Libs in 2022 due to Scott Morrison China rhetoric populations which explains some of their Liberal Underperformance.
It’s interesting seeing the realignment, especially the huge shift in Capricornia. I’ve noticed that there’s a convergence to the centre in Labor seats. Labor won lots of seats of margins of over 15%. There are far fewer Labor seats with >15% margins.
The largest erosion of Labor margins were in western Melbourne and South-Western Sydney and regional QLD (after QLDer, Kevin 07 left the scene). Gentrifying regional centres like Bendigo, Ballarat and Franklin (Tas) have seen a consolidation in the Labor 2PP.
Part of the shift in Pearce would be from redistribution – it went from being a mixed rural-urban seat to just the outer northern suburbs. Labor certainly blew the doors off in 2022 though – that would’ve been partly due to Christian Porter and the circumstances of his exit.
No, this data is after adjusting for redistributions. If you look at the chart on the Pearce guide, I’m referring to the dotted green line.
The actual Liberal 2PP in Pearce in 2007 was 59.1%, but adjusted for the 2025 boundaries was 56.2%.
Hang on, so you’re talking about the current boundaries of Pearce, but with 2007 votes? A lot of that suburbia didn’t exist back then, so your numbers might be a bit ropey. (The same quibble applies to any seat with a lot of recent suburban development.)
Pearce really is a “ship of Theseus” division. I didn’t realise it went as far south as Narrogin in 2007 – the vast majority of its area back then would be in Bullwinkel or O’Connor these days. I remember a thread on here a while back about Hume changing like that.
Sure, the current seat of Pearce has changed a lot. That’s my point.