The shift towards early voting


There has been a long term trend in Australian politics of voters shifting away from casting a ballot on election day, in favour of voting earlier in the campaign. This has usually taken place through two methods: an in-person pre-poll vote, or a postal ballot.

These trends have increased significantly in elections held since the start of the COVID-19 pandemic in early 2020.

In this post I’ve analysed the long-term trend in federal politics, and the more recent trend in state elections over the last two years.

I previously posted about the trend towards pre-poll voting in federal elections in this post back in 2019, and I’ll repost the same graph here:

Postal voting has increased, but only slowly, and it has been largely stable over the last decade. The big increase has been through in-person pre-poll voting. And that increase has been accelerating. Pre-poll voting made up 11.3% in 2010, compared to 32.7% in 2019. Only just over half of all votes in 2019 were ordinary election day votes, although for our purposes in this blog post I will be lumping in all votes cast on election day together, including absent votes, which are cast at regular booths outside of the voter's home booth.

We've seen similar trends in pretty much every state election, although the exact details vary depending on electoral procedures that either encourage or discourage the shift away from election day. There's also some variation in how election results are reported. The AEC publishes separate figures for each pre-poll centre, while some state commissions only publish total figures for each electorate. Indeed South Australia only publishes a single 'declaration votes' total for all types of non-ordinary votes, including pre-poll and postal ballots.

We've now had quite a few elections during the COVID-19 era, including four state elections, two territory elections, two federal by-elections and numerous state by-elections and local government elections. Indeed the only major electoral events not to have been held under COVID-19, apart from a federal election, are the state elections in our two biggest states.

This next chart shows the shift in how people have voted in the four state elections held since 2020, as well as the ACT election. I left off the Northern Territory election as it is quite unusual in terms of vote share.

Every state saw a noticeable decline in election day voting, but it varied. I suspect the variation can be explained by the extend of outbreak in each state at the time of the election, or at least their previous experience. The change was most mild in Tasmania and South Australia. The biggest drops were in Queensland and ACT, which held elections in the first year of the pandemic.

There's also variation in where that vote went. The ACT and Tasmania saw very little increase in postal voting, but it grew quite a lot in Western Australia. Still, every state saw pre-poll voting take on more of those voters shunning election day. Unfortunately South Australia does not break down the various non-election day voting options.

There have also been two federal by-elections since the start of the pandemic, both held in 2020 for the seats of Eden-Monaro in south-eastern NSW and Groom in Queensland. Eden-Monaro was a very close race, while Groom was an easy win for the LNP, but both featured a contest between the major parties. Conveniently for us, it also featured mostly the same rules that apply at this federal election.

The downward trend in election day voting was less dramatic, but was there - almost 10 percentage points in each seat. The increase in the pre-poll vote was quite small, but the postal vote increased more significantly in each race.

So what do we know yet about 2022?

In past federal elections, we had three weeks of pre-poll voting, which would have meant pre-poll booths opening today. Since the last election, it has been reduced to two weeks.

Pre-poll voting does not happen evenly across the voting period. The rate tends to accelerate later in the campaign, with less than a fifth cast during the first of the three weeks. You can see this in a chart in this blog post from 2019.

The push to reduce pre-poll voting may have been partly motivated by a desire to push voters back to voting on the day, but I suspect cutting a week won't do much to achieve that outcome. But it will certainly reduce the number of hours of booth work for campaign volunteers handing out how-to-votes.

While pre-poll voting hasn't started, the postal vote process is well under way.

As of last night, 1.49 million applications have been made for postal voting. This compares to 870,429 applications made at this stage in 2019, and a final number of 1.6 million applications in 2019.

Not every application results in a postal vote - those 1.6 million applications in 2019 resulted in 1.24 million postal votes.

Even still, that's a 71% increase in applications at this stage in the count. And 12,437 votes have been returned to the AEC.

All of this is consistent with a significant increase in postal voting, as was seen at the 2020 federal by-elections. i will track these numbers, and the pre-poll numbers, over the next three weeks.

Next up - I'll be back on Wednesday with a blog post about which electorates have higher rates of early voting, and then later in the week I'll have a blog post looking at how voting patterns vary between early voters and election day voters.

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


  1. I’ve never voted pre-poll, but am wondering if I am correct in the understanding that no attempt is made to ascertain whether a voter fits into one of the Schedule 2 categories?

    As I understand it, pre-poll votes require an “application” (sections 200C and 200D) and you are supposed to only be entitled to a pre-poll vote if you fit one of the reasons (section 200DG). But I’m guessing the application amounts to no more than showing up and saying you want to do a pre-poll vote, with no further enquiry other than the standard ones in section 200DI, which are also asked in relation to ordinary votes?

  2. I have my suspicion that this is being driven by the electoral commissions for financial reasons. Given they have to set up and open offices in the lead up to the election day, and the premises need to be suitable for handling large volumes of traffic anyway, why not add some staff each day in the few weeks and get a big chunk of votes out of the way? The expensive alternative in an increasing population, is to find and open more venues on the actual polling day (I’ve seen a few polling places, in my time working for both the electoral commission and now as a party hack handing out HTVs, and most could not just increase capacity in the current space they have (size of room, entrance access, parking))?

    Only one experience, but I pre-polled in a Victorian state election in the early 2000s and was given a fair “grilling” over why I couldn’t vote on the actual day, they could keep that up if they truely wanted to get the majority voting on the actual day, but seems now it’s “come on in early”

    I’m a purist, I think an election day should be actually that, a DAY, but I also get that in this environment of “customer service” we have to make things easier for all…. If that’s the case, why not admit that we now have a formal election “period”, limit it to a week and have even more centres open (maybe even the Saturday and Sunday before, and even some extended opening hours? ?), if you can’t make that week, postal votes would always still be available…

    But one thing’s for sure, these pre-poll centres should be counted and reported on the Saturday night

  3. I don’t think you’re correct Alistair. Pre-poll votes aren’t just received at the commission offices, they set up numerous centres that are stand-alone for that purpose. I think it’s a lot more expensive than just having facilities open for election day and having one pre-poll booth, as used to be the case.

    I believe most pre-poll votes will be counted on the night, but probably later on. Postal votes won’t be counted until Monday.

  4. Given the increase in pre-poll/early polling. Why don’t the pollsters make more of an effort to do exit polls like they do in the States and UK?
    Is this because Aussies are a little bit more cagey about letting people know who they voted for? Or because it is compulsory and many are not there because they want to be but because they have to be.

  5. Election night would be better if the AEC were able to count prepoll (and postal votes received before election day) on election day before polls close. Security and secrecy provisions probably make it too hard.

    I’d rather wait a long time for results than have FPTP or disenfranchised voters

  6. Hi Ben – I might be missing something…you mentioned “‘I’ve also mapped out the electoral map, showing which seats had higher or lower proportions voting early in 2019.” but I can’t find it?

  7. WRT volunteers manning prepoll booths: my experience is that the voters who get in the earliest are the least persuadable anyway. It’s not just a convenience thing, they know who they want to vote for and want to do it as soon as possible.

  8. I agree Furtive (!) – pre-poll voters def seem to have their minds made up & just want to get it done. Depending on the seat & the capacity of the volunteers, postal votes seem to be where a slick operation could harvest votes. Sitting here in Higgins wondering if it’s going to be an interesting night… or days.

  9. I can not see why for all elections After the 21 May 2022 election pre poll votes can not be counted on Election Day within a secure environment And prepoll count released 15 minutes after poll closes. Then again I can not see why postal votes must be dealt with in the same manner. Political parties Manage to get postal vote applications out within hours of the writs being issued.

  10. Can confirm the last couple of elections they haven’t pushed voters too hard on the reasons why they were voting early. At least not in my area (Carrum Downs/Frankston in Victoria).

  11. I have suspected that it suits the AEC to have people vote prepoll. In 2010, I had to queue for 45 minutes on polling day – queues I had never seen before or since. And this was my local polling place. Did AEC understaffing piss off enough folks to vote prepoll in future?

  12. Andrew J.
    I appreciate your reasoning and agree that they could be locked in a room from early in the day. However, announcing at 6:15 !! You sir are a psephological killjoy!! Could be all over bar the shouting – no drama, no Antony Green having tech problems, no politicians being nice to each other (happens once every three years). And worst of all – no pontificating from the couch – my pontificating that is!! However, a huge drop of prepolls results at say 8pm would be quite exciting!!

  13. Unemployment figures are released on the Thursday before the election and are expected to be the lowest since Whitlam. Prepoll voters will not be influenced by the number.

  14. Menzies came within a whisker of losing the 1961 election because unemployment rose above 1% . How standards have dropped.

  15. I doubt anyone at all will be influenced by that number. What does 4% unemployment matter if the people who *do* have a job can’t pay the bills? Seeing as people working as little as 1-2 hours a week are counted as being “employed”, I don’t think the number means much to the average punter.

    People who vote prepoll are also the people who already knew months ago who they were voting for anyway, the timing will make no difference whatsoever to them.

  16. Have you seen any statistics yet on the sources of postal vote applications in 2022? It would be especially interesting to know how many are made on-line directly with the AEC, compared with those which come via political parties. (I have a strong aversion to the parties’ role in this, accentuated by the way in which parties thereby get access to the answer to the voter’s “secret question”, the use of which is presumably intended to be an integrity measure.)

  17. Michael, the AEC don’t publish that data anymore. That’s a change since 2019, they DMed me on Twitter to explain the decision:

    “This data will include a breakdown by electoral division. However, for this federal election it will not include a breakdown regarding the source of postal vote applications. This is in part due to the digital evolution in how political parties are distributing postal vote applications, resulting in applications that are unclear whether or not they have come from a third party source. It is also a result of the expected volume of applications and how our election management systems are able to extract the data.”

    They are also not publishing daily totals retrospectively. I’m trying to remember to download the data each night to get a trendline but have missed a few days.

Comments are closed.