Prior to the election, I published my own estimates of who would have likely won seats in the Victorian Legislative Council if group voting tickets were abolished and replaced by a system similar to those used for the Senate or for other upper houses in NSW, Western Australia and South Australia.
I’ve done a similar exercise on the 2022 results, but also followed up with estimates of how the results would have looked under various other models.
This chart shows the seat count for each party under the various models:
I've identified seven seats which I think would have changed hands in the absence of GVTs. The Greens would have gained two seats at the expense of Legalise Cannabis. Labor would have gained a seat at the expense of Animal Justice and the Liberal Party one at the expense of the Liberal Democrats.
In the case of the Democratic Labour Party and the Shooters, Fishers and Farmers, I think they would have missed out on one seat, and gained a different seat, although one of these cases is the hardest to judge.
These are the seven contests across five regions:
- Eastern Victoria - The hardest one to call. In reality this seat went to the Shooters, Fishers and Farmers but I picked it as going to the DLP. The Greens were leading for the fifth seat with 0.48 quotas, but there were a lot more right-wing votes remaining in the count than left-wing votes, with Labor on track to win the fourth seat with just 0.59 quotas. The seat did end up going to the Shooters, Farmers and Fishers, but the Democratic Labour Party had a higher primary vote at 0.23 quotas, so I've given the seat to them. But you could argue the SFF would have won this seat without GVTs and thus it wouldn't be on this list.
- Northern Metropolitan - The DLP's Adem Somyurek won the fifth seat. While the DLP did poll quite well with 4.8%, more than any of the other parties competing for that seat (admittedly with the help of a confusing name and a strong ballot position), there were three strong left-wing minor parties not far behind. I think the Victorian Socialists would have been most likely to win.
- Northern Victoria - The last two seats went to Animal Justice and One Nation. Without GVTs I think these seats would have gone to Labor and the Shooters, Fishers and Farmers.
- South Eastern Metropolitan - The last two seats went to Legalise Cannabis and the Liberal Democrats. I think they would have instead gone to the Greens and the Liberal Party.
- Western Metropolitan - The last seat went to Legalise Cannabis, but instead I think it would have gone to the Greens.
The upper house without group voting tickets would have roughly the same political balance, but dominated by a smaller number of parties. With the exception of the Victorian Socialists replacing the DLP, every other seat change involves parties on the left replacing parties on the left, and the same on the right.
Labor, the Coalition and the Greens gain four seats, with Animal Justice, Reason, the Liberal Democrats, One Nation and Legalise Cannabis knocked out of parliament. Just three small parties remain.
The number of seats changing hands is down from the nine in 2018, which reflects the change in preference strategy, with minor left parties preferencing the Greens and Labor before the other minor parties - there were a number of races where a Druery-aligned minor party could have won if parties like Legalise Cannabis, Reason and Animal Justice favoured them, but instead went to Labor or Greens.
While I'm considering alternative models, I also repeated my exercise of looking at who would have won if every vote was above the line using group voting tickets, by looking at who was elected under Antony Green's preference calculators. Overall the below the line rate increased slightly from 8.8% to 9.5%.
Two seats would have changed hands without below the line voting - both would have involved the Greens losing a seat to a party broadly on the left. The Greens won in North Eastern Metropolitan, defeating Transport Matters, and also won in Western Victoria, defeating Legalise Cannabis. The smaller party in both races had more favourable preference flows and would have won if every vote followed the preference ticket.
Finally, I wanted to look at what would have happened if the entire Council was elected as a single electorate of 40 MLCs with no regions. I advocated this system before the election, and I note that the newly-elected Animal Justice MLC has argued for the same.
In the first table further up in this post, you can see the results for three counting systems - under the same Single Transferable Vote (STV) system we use for the NSW Legislative Council (which is also effectively largest remainder). I then looked at two divisor systems which don't involve preferences are much easier to count. Under D'Hondt, you divide the party's vote by the number of seats they've won, plus one (1, 2, 3, 4), while under Sainte-Laguë you divide by the number of seats multiplied by two, plus one (1, 3, 5, 7). It turns out STV and Saint-Lague produce the same result, so I've collapsed them into one column.
All of these models elect more minor parties than under the M5 regional system without GVTs. D'Hondt produces a very similar outcome to the actual result, but with one more for Labor and the Greens, one less for Legalise Cannabis, and none for Animal Justice.
Saint-Lague and STV at large produce an even more diverse crossbench, with the major parties only winning 25 seats between them along with 4 Greens. Legalise Cannabis win their two seats as the highest-polling small party, and another nine parties win a single seat.
The change in preference strategy means that there aren't such large divergences between the actual results and the statewide outcomes, although of course we know the current system still has the potential for that divergence unless it is fixed.
This prompted me to actually calculate the Gallagher index of disproportionality for upper house elections too. I calculated the lower house scores the other day, finding that the 2022 election was the most disproportionate since the 1960s.
I haven't gone quite so far back for the upper house, collecting data since 1996. This chart shows the score for both houses over that time.
The old Legislative Council used to be a bit more disproportionate, which makes sense considering it used the same system but with a quarter of the electorates. But when the Council was reformed, the score dropped significantly. And while disproportionality did increase to a peak in 2014, the 2022 election is the least disproportionate upper house election in the last three decades (and likely so for the entire history of the upper house). Something worth remembering when we focus on the ills of group voting tickets - ultimately the lower house is worse.
