One way of looking at the Victorian election is to examine which seats ended up being ‘non-classic’ contests – seats where the two-candidate-preferred count didn’t end up being a Labor vs Coalition count.
This statistic peaked in 2018, and has gone slightly backwards in 2022, but the number is still much higher than in pre-2018 elections.
After all the hype about independent contests, the number of non-classic contests involving independents ended up dropping in 2022, but the number of Greens contests increased.
The Victorian Electoral Commission doesn’t provide the same completeness of data as the AEC. While the AEC conducts a complete distribution of preferences in all seats, the VEC only distributes preferences as far as is necessary for a candidate to pass 50%. So in a couple of cases where it’s not clear I’ve had to judge who would have made the final count. I’ve included the independent in Mulgrave in 2022, and the Greens in Footscray and Preston in 2018.
The VEC does conduct a notional two-candidate-preferred count in every seat, but they don’t bother to redo the count if the distribution of preferences ends up producing a different pairing.
The number of non-classic contests first jumped in 1999, with 6-8 contests at elections from 1999 to 2014.
The number surged in 2018, with 14 non-classic contests, but it's now dropped back slightly to 13.
When you look at each category, you can see that this trend has been driven in large part by the presence of the Greens. The Greens first broke through to the 2CP in four seats in 2002, and by 2022 they've reached that count in eight seats, all in a contiguous area in central Melbourne.
The number of independent contests peaked in 2018, with seven independents making the 2CP.
The specific list of seats also gives a sense of the trend. Once the Greens have first broken through to the 2CP in a seat, they've maintained that position at every subsequent election, breaking through in Prahran in 2014, Preston and Footscray in 2018, and Pascoe Vale in 2022.
For independent seats, the only overlaps were in three northern Victorian seats: Benambra, Mildura and Shepparton. Independents made the 2CP in Geelong, Morwell, Pascoe Vale and Werribee in 2018, but not again in 2022.
An independent broke through to the 2CP in Mornington in 2022. Early on it appeared that independents were winning in Hawthorn and Kew. Not only did they not win, but they didn't end up making it to the final count.
I've also counted Mulgrave as a non-classic contest. Daniel Andrews polled a majority of the primary vote, so there is no distribution of preferences, but the independent Ian Cook did outpoll the Liberal candidate. Apparently the VEC will at some point next year conduct a full distribution which will answer this question.
As a proportion of the total number of seats, Victorian politics is not that far behind federal politics, where 17.9% of seats were non-classic at the recent federal election. The Victorian figure this year was 14.8%.
This list doesn't factor in a number of seats where independents polled strongly but didn't make the final count, specifically Hawthorn and Kew. An independent also made waves in Preston, but ultimately polled fourth.
While there are no independents in the new Victorian lower house, the recent state election was still a significant moment for independents - in at least eight seats they were competitive. I suspect this dynamic will remain for some time to come. As long as there are numerous serious independent challenges, there is the potential that different political circumstances could see the kind of breakthrough we saw at the federal election.
This is my last post about the Victorian election for a little while. I'm planning to put together my dataset of the 2022 results over the coming break and I'll be back with a bit more analysis made possible by that dataset, before switching to the NSW state election. But until next year - thanks for joining me here at the Tally Room.
Thanks for the analysis as always Ben.
I think one thing missing from this piece is looking at how some seats once they become non-classic contests have a chance of staying that way, and may alternate between parties/IND for the 2CP. While looking at overall allegiance, it sometimes misses the longer term trends. You rightly mentioned Shepparton and Mildura as staying in non-classic for IND across 2018 and 2022. However, this misses the trend that both have been non-classic since 1999 and 1996 respectively. For the record:
Shepparton 99 NATvsIND, 02 NATvsLIB, 06 NATvsLIB, 10 NATvsCA(nowAFP) that’s the ‘Other Party’ mentioned in the graph in 2010, 14 INDvsNAT, 18 INDvsLIB, 22 NATvsIND
Mildura 96 INDvsLIB, 99 INDvsLIB, 02 INDvsNAT, 06 NATvsIND, 10 NATvsIND, 14 NATvsIND, 18 INDvsNAT, 22 NATvsIND
So when you look at VIC 2014 Election, these two seats (Mildura & Shepparton) + GRNs 2CP seats explain all the non-classic seats at that election. Then VIC 2010 Election was those two seats + GRNs 2CP seats and Gippsland East. Interestingly, Gippsland East was non-classic 99,02,06,10 thanks to an IND member but has now stayed as a classic contest seat.
But you make a good point though, once GRNs get in 2CP (city/urban seats), they do seem hard to shake off, making those seats continual non-classic contests and a challenge for ALP where they have to fight on essentially two fronts.
A few seats come to mind in other jurisdictions. QLD for example:
Callide 12 LNPvsKAP, 15LNPvsPUP, 17 ONPvsPUP, (although 20 was Classic due to right wing minors not contesting)
Noosa 12 LNPvsGRN, 15 LNPvsGRN, 17 INDvsLNP, 20 INDvsLNP. Noosa is an interesting case of GRN starting in 2CP then changing to IND.
Lockyer is susceptible with most elections being non-classic 2CP but occasionally goes classic.
Hinchinbrook, Traeger (ex Mt Isa), Hill (ex Dalrymple, ex Tablelands), Gympie and Warrego are other seats that are susceptible over the long term.
FED for example:
VIC
Cooper (ex Batman) GRN since 2010
Melbourne GRN since 2007
Wills GRN since 2010
Interesting that VIC seats hold to GRN at State and Fed level once they switch.
QLD & TAS
Kennedy since 2001, Denison (ex Clark) since 2010 but both of these are due to sitting members and were classics beforehand.
NSW
New England 2001-2019 IND, went classic first time in a while in 2022.
Cowper since 2016 IND
Grayndler since 2016 GRN
Warringah since 2016. Was LNPvsGRN in 16, before 19&22 INDvsLNP. Interestingly 2001 saw this seat go LNPvsIND.
SA
Mayo since 2016 CA (ex NXT) and had a non-classic result in 1998 LNPvsDEM.
Thought that might be a comment with food for thought about how which party/IND runs and in which seat can increase the chances of that seat being a non-classic contest. I’ll look at NSW later.
Thanks for that, I did actually delve into that a bit more in my pre-election blog post on this topic and didn’t want to double up too much.
@Ben – that’s right you did! So many good posts and comments to keep up with! It’s just interesting to look back on an extra data point (VIC 2022) and compare. I found it interesting on a couple of points:
-The Greens really solidifying their position in 2CP contests at State and Fed level in VIC.
-With the slow decline of the Major Party vote, a collapse in certain areas has lead to higher 2CP contests. Simplified (we can delve into lots more detail for another time) LIBs loosing support in VIC has lead to rise in GRN in 2CP (especially inner-city Melb) while LABs loosing support in NSW has done the same for GRNs (Northern Beaches & North Shore area comes to mind) and INDs (Tamworth/Northern Tablelands and a couple of other regional areas come to mind).
To illustrate the collapse of the ALP vote in 2011 NSW, GRNs finished in 2CP in 14!!! non-classic contests, but only 11 in 2015 and 7 in 2019.
-The GRNs really have positioned themselves to be in 2CP contests at many by-elections when there isn’t a strong IND and as the alternate when a Major Party isn’t running. VIC doesn’t really have a strong party to take up when LIB isn’t contesting a by-election (Family First-2013 Lyndhurst an exception) while NSW did have the CDP (2016 Canterbury, 2017 Blacktown) but they no longer exist. Until a strong ‘right-wing’ party comes up, the GRNs will just get more chances to get into 2CP and cause possible more upsets without strong competition.
Your bar graph does show nicely that it’s a slow build in VIC and would most likely increase at the 2026. Unlike QLD which is susceptible to having huge flash in the pan non-classic contests like KAP first starting out in 2012 and 2017 thanks to a resurgent ONP, only to fizz out again in 2020 QLD. To demonstrate this, the non classic contests for QLD were 6 in 2009, 18 in 2012, 7 in 2015, 28!! in 2017 and 8 in 2020. With NSW to compare, non-classic contests were 21 in 2007, 27!! in 2011, 16 in 2015, 19 in 2019.
I don’t buy that the contests went backwards. Help this make sense to me. I know Geelong lost it’s non-classic contest. But which other seats in 2018 didn’t make it this time? I find it hard to believe.
There are also a bunch of teal target seats where the full distribution of preferences are not finished.
Daniel
As well as;- Morwell, Pascoe Vale and Werribee
Pascoe Vale became a Labor vs Greens seat at this election. So that’s a gain to the non-classic contest. Morwell lost it’s status.
We still don’t know the final count in Hawthorn as in who will finish 2nd after preferences, (Lowe still has a small chance of making it 2nd after preferences) I would also throw a few other seats into the “too early”
Which seats am I missing? There are more Green vs ALP seats this time around as well as more independents considering while the Independents in the rural area’s lost re-election, they still made the final count.
A map showing the non-classic contests would be useful.
Pascoe Vale was non-classic in 2018, it just switched from independent to Green.
The subtractions from the list this year:
-Geelong
-Morwell
-Werribee
The additions:
-Mornington
-Mulgrave
The seats on the list both elections:
-Benambra
-Brunswick
-Melbourne
-Mildura
-Northcote
-Pascoe Vale
-Prahran
-Preston
-Richmond
-Shepparton
-Footscray
I did think about making a map but it fell by the wayside with the Christmas rush.
Ben, so that means Hawthorn was a traditional ALP v LIB 2PP contest. Did the VEC conduct the full distribution of preferences to reach the final 2 candidates or did it stop at the 3CP level (Liberal, Labor and Melissa Lowe).
It seems Hawthorn this year is like the contest in Hartley, SA at 2018 state election featuring three high profile candidates – namely incumbent Liberal MP Vincent Tarzia, Labor candidate Grace Portolesi (a former MP herself) and SA BEST candidate Nick Xenophon (a former Senator and party leader). Eventually the final 2CP count ended up as a traditional ALP v LIB contest.
The anti-preferential voting, pro-FPTP crowd are normally very vocal when Labor wins an election but they were very silent at this election.
It’s probably because the Nationals gained seats with Liberal preferences. It might also be because the minor parties like DLP, Freedom, Family First and LDP split the conservative vote and so Labor came first on primaries in most seats. I mentioned this in another thread before.
The Greens actually first came second (in the 3PP) in Preston in the 2014 Victorian State election rather than in 2018 – but the VEC didn’t do a full 2CP count because the Labor MP ended up with over 50% of the vote on the 3PP count:
https://www.vec.vic.gov.au/results/state-election-results/2014-state-election
@Yoh An, a full distribution was done in Hawthorn because Pesutto didn’t pass 50% until the final count.
It wasn’t until the full distribution that the result flipped from LIB v IND to LIB v ALP but the result was much the same.
In Brighton, however, no full distribution was done (yet – they are being done in February) because Newbury passed 50% in the 3CP count.
@ Trent, do you know if Caulfield, Sandringham and Kew had a full preference distribution?
Caulfield & Kew did, but Sandringham didn’t because Brad Rowswell got to 50% in the 3CP count.
Here were the 3CP counts of the 3 which had full distributions:
CAULFIELD:
LIB – 48.4%
ALP – 30.7%
GRN – 20.9%
(Then the LIB v ALP 2CP was 52.1-47.9)
HAWTHORN:
LIB – 45.3%
ALP – 27.5%
IND – 27.2%
(Then the LIB v ALP 2CP was 51.7-48.3)
KEW:
LIB – 46.9%
ALP – 26.8%
IND – 26.3%
(Then the LIB v ALP 2CP was 54.0-46.0).
Here were the 3CP counts in Brighton & Sandringham which never went further:
BRIGHTON:
LIB – 50.3%
ALP – 28.8%
GRN – 20.9%
SANDRINGHAM:
LIB – 51.2%
ALP – 29.2%
GRN – 19.6%
The “teals” only even made it to the 3CP count in Kew & Hawthorn. They were eliminated at the 4CP count in Brighton (10.7%), Sandringham (8.3%) and Caulfield (7.5%), finishing well behind the Greens in all three.
Thanks Trent for the information. I am still shocked that in Kew and Hawthorn the teal did not make the final 2CP even though both AJP and Greens preferenced the Teal ahead of Labor. Also do you feel in Caulfield and Brighton the Teal running actually helped the Libs by taking away Soft Labor votes and some of it returning to Libs in preferences?
I feel like Labor could have retained Hawthorn if the teal hadn’t run. She split the progressive vote and some went back to the Libs. Not saying Labor would’ve won, but rather they could’ve. Still may have been hard as the mega rich areas of Melbourne like Brighton, Malvern, Nepean all had swings against Labor unlike the more middle class seats further east.
Comments are closed.