Tally Room mail - how might Victoria play out? (Patreon)

It's been a while since I wrote my Patreon newsletter, but we are now in a quiet period for Australian elections and I won't necessarily have as much to write on the website for the next while. So I'm going to try and write regular newsletters to keep you updated on my work and give my thoughts.

I'm going to publish this for $8+ patrons, and then unlock it for everyone else at the end of the week.

What might happen in Victoria?

There's so much that's unprecedented about One Nation's surge in support nationally. We know that there won't be a straightforward relationship between the total polling figures at a state or national level and who wins seats. We are learning more about where One Nation is likely to do well, and how preferences may flow, but there's more to know.

There is quite some time before the next federal election, and it is possible the One Nation surge may turn around, but right now it looks pretty real, and they had no problem carrying out their campaign in the recent elections in South Australia and Farrer. Some people prefer to live in denial and use wishful thinking to reassure themselves, but One Nation really is now polling close to 30%. That's probably not enough for them to form government, but there's still a long way to go.

The MRP polls that have been conducted federally are very useful. There was a Demos MRP early this year and a Redbridge MRP in May, both of which showed similar figures. But since that last Redbridge MRP we've seen One Nation again climb in the polls after a few months of stable polls. I think that has given us a bit more of a sense of how the current polls would play out in federal politics. But I'm right now more interested in how One Nation support may play out in the Victorian election, due in November.

So far we've only seen One Nation's surge play out in elections where the Coalition or Liberal Party is at a low point. The South Australian election was already looking like a Labor landslide before One Nation entered the fray. But that is not the case for Victoria.

Recent polling suggests that both Labor and the Coalition have lost support in Victoria, but it's likely that there has also been some swing from Labor to Liberal in the cities, which makes the Coalition-One Nation swing look more modest.

We've had two Victorian state polls since the Nepean by-election. Freshwater reported 27% for the Coalition, 25% for One Nation, 23% for Labor and 14% for the Greens. Demos reported 30% for the Coalition, 23% for One Nation, 21% for Labor and 15% for the Greens.

Unfortunately we don't have any MRPs for Victoria, so I decided to do a back-of-the-envelope estimate of how one of those polls might play out. I'm not going to publish the actual figures, or how particular seats played out, because it's basically an educated guess. But I think it still has some insights to offer.

We know that One Nation's support is strongest in rural areas, but also on the outskirts of suburbia. I would expect them to peel Coalition support in rural areas and Labor support more in the outer suburbs, and I assumed a similar pattern to One Nation's upper house support in 2022. I also assumed the Greens gaining some ground from Labor in the inner city, and a uniform swing making up the remaining difference from Labor to Coalition. I applied those assumptions to the 2022 results, which obviously misses any local nuance, but gives us a rough sense of how many seats you'd expect parties to win in such a scenario.

Based on the Freshwater poll, I ended up with Labor, the Coalition and One Nation all with 20-something seats each. The Coalition and One Nation had enough seats for a majority between them, but so did the Coalition and Labor. Labor came third on the primary vote, but still won the most seats. This reflects that Labor still would do relatively well on Greens preferences, although the Greens expanded their seats with such a high vote.

I genuinely have no idea how such a parliament would play out if it were to occur. A minority Coalition government could govern with the tolerance of Labor, One Nation or both. But the right-wing majority wouldn't be large and could be lost if either side loses some of their numbers. Alternatively, this is a scenario where I could imagine Labor and Coalition working together in some way to keep One Nation away from influence - but I can't imagine such an outcome would be bad for One Nation's polling.

And hypothetical scenarios like this one will undoubtedly be discussed during the campaign. The Coalition is hoping to gain votes from Labor in areas where One Nation is weak. Will those swing voters hesitate if a Coalition government will be reliant on One Nation? I don't think anyone can know how this will play out.

Compiling more datasets

A big part of my psephological work is compiling datasets for election results that are tidy and easy to analyse. That then means that, during busy election periods, I'm able to pull together analysis very quickly instead of spending all day wading through results websites that aren't in any easy-to-use format.

I'm currently finishing a project I started during the summer - a dataset of federal election results all the way back to 1901.

The House results are almost done - I think for now I'm going to leave the Senate as a separate project to do later. It includes seat-level results back to 1901, and includes two-candidate-preferred and two-party-preferred figures where appropriate. It includes primary booth-level results back to 1993 and 2CP/2PP booth results back to 1996.

There were some tricky dilemmas when it came to estimates of 2CP before 1984. Prior to 1984, the AEC would only distribute preferences as far as necessary, so if a candidate achieved a majority prior to the final count, there would be no final count. A number of analysts produced their own estimates of the likely 2CP in those cases, and those numbers have made their way into current online results archives like Wikipedia and Psephos, but it's not always clear who made the original estimate. I have tried to identify the original source, but that's a continuing project.

I am now adding data from the 2004-2025 era, during which the AEC results are reasonably clean and easy to put together. But I expect in the next week or so I'll be able to add this data to the data repository for patrons to access, along with a seat-level results archive for Tasmania dating back to 1989.

Right now I've got datasets at the seat and booth level for pretty much all elections back to about 2010-11, but I will be planning to do much deeper seat-level historical datasets for various states. I have also been attempting to compile datasets of historic 3CP results, which I have begun adding to the appropriate data repositories. Hopefully over time we will have more complete data that is easier to find.

Access to journals

I have a research note that will be published soon in the Australian Journal of Political Science entitled "Rule of three: the growing relevance of three-candidate-preferred voting counts in Australian elections". Unfortunately I have discovered that it won't be published open access.

I am not an academic, but I do now have a visiting fellowship at the Australian National University. It has been an enriching experience to write a journal article, but I think in the future if my work is going to be locked away without open access, I would prefer to write for the blog, or even look for other writing instead.

Anyway, look out for the article! I plan to share a version of the article with Patreon subscribers when it is finished.