This blog post is an update to two blog posts I wrote in 2022, looking at the rate of counting progress on election night and how the primary vote for Labor, the Coalition and the Greens shifts over the course of the night.
First up, this chart shows the proportion of the total votes cast that were counted (as House primary votes) at each minute of election night, for the eight states and territories.
If you compare this to the equivalent 2019 chart, most states show a much slower count in 2022 than 2019.
We can compare these more directly by merging the data into a single national chart for House primary vote counting progress, and comparing it to the same line for the previous four federal elections.
There is a notable slowing down of vote-counting at each election. At 8pm on election night in 2010, almost 45% of House primary votes had been counted. Only about 40% had been counted by that point in 2013, and around 30% in 2016 and 2019. Only about 21% of the vote had been counted by 8pm in 2022.
This change in counting trends means that the bulk of counting stretches out deeper into the night, rather than having a rapid count early on and then little additions later in the night.
In 2010, 77% of the vote was cast by 10pm, which was almost all of the votes counted by the end of the night. This number dropped to just over 60% by 10pm in 2019, but the final vote share counted by the end of the night ended up in a similar place, around 80-84% of the final vote totals.
In 2022, however, significantly fewer votes had been counted by the end of the night, reaching 50% by 10pm and not quite reaching three quarters of all votes counted by the time counting stopped around 2am. I think this was due to a combination of an increased number of postal votes that weren’t counted on the night, and some pre-poll booths not reporting primary votes on election night.
After initially publishing this post I have added an extra chart which I think also illustrates this point. It shows the raw number of votes added per hour – not a cumulative count.
Over 5 million primary votes were added to the count in the 7-8 hour in 2010, but it was only 2.9 million in 2022.
In contrast 2019 and 2022 recorded more votes later in the night, but it doesn’t fully compensate for that slow early start.
So far I have been entirely focused on the counting of House primary votes but there are three different kinds of votes counted on the night: House primary votes, House two-candidate-preferred counts, and Senate primary votes.
Unsurprisingly, House primary votes and then House 2CP votes make more progress with counting than Senate primary votes. But the levels of progress are interesting.
While the 2CP count is usually about 45 minutes behind the House primary count, by the end of the night it isn’t that far behind. 74% of primary votes are counted, along with almost 71% of 2CP votes.
The Senate count is substantially less progressed, with the count flattening out just short of 50%.
If there is an increase in pre-poll voting in 2025 and postal voting remains steady, I don’t see these trends changing – it is still taking longer for votes to be counted, and this means that we are likely to have a less complete picture at the end of the night.
Finally I’m not going to provide much analysis of the final three charts, showing the primary vote for Labor, the Coalition and the Greens throughout the night at the last five federal elections, but they give you a sense of how long it takes for the vote counts to resemble the final results.
The most interesting element is how Labor’s primary vote in 2022 did not follow its previous shape. Instead, Labor’s final primary vote was significantly understated until votes were reported from Western Australia, where Labor did much better than they have in the past.
2019 was the first Election where Prepoll was counted on the night, rather than the following Tuesday. From memory, counting 14 ,000 votes and doing the 2CP took until 10:30 pm, whereas counting HoR at a medium size booth of 2,000 would be done by 7:45 pm at previous elections.
Working as a scrutineer for the past 6 elections in a small regional booth, the expertise of the temporary AEC staff seems to have declined over the years.The staff change each election, most are newbies and don’t appear to use staff to efficiently divide up work so that the counts get under way quickly e.g they go step by step through instruction booklet rather than doing two steps at once with staff available ,meanting that some stand round doing nothing before getting on with the count.
And they can follow the book too closely: one election the green candidate in our revolutionary village polled more primaries than the (sitting) Liberal,behind Labor. The AEC boss wanted to distribute Lib preferencs, but I and others assured them that our booth was unrepresenative and that they should check further up the chain of command before wasting their and our time with an aberrant count.
@megan stoyles the declining standard of AEC staff over time reflects declining standards of literacy, numeracy, and average IQ in society generally.
@Megan Stoyles – isn’t the 2PP count electorate wide rather than specific to each polling place? They were correct in distributing to the predetermined top two candidates unless it was an electorate wide trend.
Megan Stoyles
The on-the-night 2CP count is merely indicative. The AEC will have a predetermined pair of candidates that they will presume as the final 2 candidates. However this is never final, they always do a real Distribution of Preferences later with the proper sequential elimination of candidates.
I think Megan understands. She’s saying the AEC official wanted to run the indicative 2CP between Labor and Greens even though that wasn’t their instructions.
Is the general thinking that the more early voting there is then the longer it will take to get a clear winner on election night?
I am saving my vote till Saturday in the hope that my vote is counted sooner
@ Ben Raue
Oh too right, silly me. I had what Megan said upside down in my mind. That is quite a rookie mistake from the AEC official.
I agree with Megan – having both worked as an AEC official in my youth and run a a booth at a council election in my 20’s, I have now scrutineered a bit at recent state and federal elections. They are soooo SLOW…
First of all they have a 20 – 30 minute break, then they all stand around waiting to be told what to do. It generally takes till 8pm to get the first preference count for a mid sized city booth of say 1,500 votes. I just checked – it was 7:54pm when I took the photo of the primary vote tally in 2022 and it would have been some while after that before it made it into the AEC feed.
I won’t do it again unless absolutely have to – it is too frustrating
Presumably that one-off early strong start for the Greens in 2022 was the Norfolk Island booth reporting before 6pm. I believe that may have been done accidentally, but I hope not — it was very helpful in confirming that election results systems were working. Plus it would be nice to have our very own Dixville Notch.
@Megan, I understand and agree with what you’re saying. Part of the problem, I think, up until at least 2016, was that the wage rates per hour hadn’t shifted in a long time, and was frankly uncompetitive. It meant that fewer applicants, more turnover, and increasing pressure on the people who did turn up for work.
The hourly rate did improve, but I am not sure that it improved by enough each election to stay competitive – and likely not by enough to attract people back to do it again. Put that together with demographic changes and in some locations there are just not enough people full stop in quite a large radius.
I think it’s going to need some “out-of-the-box” thinking in future to solve it.
The AEC got asked about Norfolk Island in today’s media briefing and they basically said even if the results from Norfolk Island are ready before 6pm they will try to wait until 6pm to publish it in the results to avoid the confusion of results going up before 6pm. So possible it will just go up right on 6pm.
Thanks – better than nothing. Congratulations on accessing the AEC media briefing. I tried and failed.
I hope that the AEC decides to do a 3CP count for the ones where it will be more useful – Macnamara, Richmond, probably a few others – than taking a wild guess on who’s making final 2.
They won’t be doing any 3CP counts on the night. They may do one later if need be.
With the clash of prepolling with public holidays, school holidays for many and annual leave, there may be more diversity in people voting as the time off is like their normal ‘Saturday’. With more prepoll and postal voters and more diversity, this may create a ‘blue mirage’ – an effect where the results are more skewed to the Coalition earlier in the night (when looking at AEC raw figures). The votes of those at larger polling booths will get counted later.
Ben – where did you get the vote count by the minute after 6pm? AEC’s media feed?
The opposite of the Norfolk Island question is Parkes having two time zones (a different one for Broken Hill). The AEC may wait till after 6.30pm AEST to publish?
Yes, the media feed.