The mismatch in Brisbane’s booth results

When I was analysing the 2020 Brisbane City election results at the start of the year, I noticed an inconsistency. I’ve now identified that the same issue is true of the 2024 results, and I wanted to clarify what is going on.

This issue only affects the lord mayoral results. In short, there’s approximately 100,000 votes (about 14.2% of the total) which have been reported in the incorrect ward. This is because votes cast outside of a voter’s home ward have simply been counted at that booth, rather than being counted as absent votes in the home ward.

Firstly, it’s worth clarifying the typical practice in Australian elections.

Australia makes it easy for someone to cast a vote, wherever they are. In many other countries, there is a single polling place where each voter can cast their ballot. But here, a voter can cast an ordinary vote at any booth in their electorate. Usually voters can also cast an absent vote outside of their electorate. At federal elections, you can cast an absent vote anywhere in their home state.

In the City of Brisbane, you can vote anywhere in the city. If you vote within your home ward, your vote counts as an ordinary vote. If you vote outside of your home ward, it’s an absent vote.

This requires an extra step in the counting process. Absent votes need to be piled up and transported to the correct electorate, where they can be counted.

For the 2020 and 2024 Brisbane City elections, this process was followed for the council ballot. But for the mayoral election, it doesn’t look like this was done. I was led to believe the 2020 process was modified as a COVID-19 measure to minimise movement of ballot papers, but the results suggest it was repeated this year.

So how many votes were actually affected? To do this, I started by looking at the number of votes cast on the two ballot papers in each vote category.

I’ve grouped the votes into four broader categories: in-person early voting, in-person election day voting, Brisbane City Hall votes, and other categories.

It makes sense that you’d see small changes between the categories, since there might be places where one ballot is counted but not the other one.

It’s pretty obvious when you look at this table. Not a single vote was recorded in the absent vote categories, either on election day or at pre-poll. Those two categories added up to about 100,000 votes on the council ballot.

Those votes didn’t disappear, but appear to have been credited to in-ward polling places, for roughly the same number of votes.

Unfortunately that means that this data inconsistency doesn’t just infect the ward-level results, but also affects each individual booth. Almost every booth in the city reported more votes for mayor than council – on average the mayoral ballot reported 19.8% more votes than for council.

And if you look at the ward-level results, you can see that the exchange of votes wasn’t even between each ward. Some wards had a much higher turnout on the mayoral ballot than in others.

Calamvale, Doboy and McDowall all had a lot of votes counted in other wards, while Bracken Ridge, Forest Lake and Wynnum-Manly seem to include votes that really belonged in other wards.

I really don’t think this is ideal, and the ECQ should make sure they don’t use this practice again at the next election. While all votes have ended up counting towards the correct result, there is value in ensuring accuracy of vote counts. And they are already transporting council ballot papers, so I don’t see why they can’t do the same for the other ballot.

Despite this issue, I do think it is still worth using this data for analysis. About 14% of votes have been swapped between wards, but that still means the vast majority of votes are local. But it means that the figures are not as accurate as they could be.

Liked it? Take a second to support the Tally Room on Patreon!