Speculation about an early federal election never goes away, and I’ve noticed more recently in relation to the possibility of a federal election later this year.
While you can debate the political merits of such a decision, the current redistribution processes make it pretty much impossible to call an election in 2024, without triggering special mini-redistribution procedures which would produce severe inequity in electoral populations between seats.
In this post I will try to understand what is going on. Is this just the normal process, or is this redistribution actually taking a long time? Yes, in short, this year’s procedures have been taking longer. I’m going to look at the time frames for past redistributions and how the timing of redistribution procedures interact withe election time rules imposed by the constitution.
First, let’s try and understand some basic rules of Australian election timing.
The House of Representatives and the Senate technically follow two different schedules. The Senate has a fixed term, with terms ending every three years at the end of June. The current Senate terms expire on 30 June 2025 and 30 June 2028. A half-Senate election can be held at any time in the last year of the Senate’s term, but the new Senate doesn’t take over until the new term.
The House, on the other hand, does not have a fixed term. Theoretically it can be dissolved at any time, but expires three years after the first sitting of the House after the last election.
In practice we always hold these elections at the same time – the last time the elections were held apart was in 1972 – so the range of possible election dates is dictated by the overlap of the two schedules.
A half-Senate election can be held from any time in early August 2024, with the final dates in May 2025. A House election can be held up until 27 September 2025. Because the election is due to be held in the early part of the year, close to the end of the Senate term, there is a wider range of potential dates.
When elections were held in the August-November window from 1998 until 2013, the end of the House term imposed a much tighter window. This schedule was reset by the double dissolution in 2016.
But the other constraint is the process of redrawing electoral boundaries.
One year after the first sitting of the House, the entitlement of seats in the House for each state is determined. Once the determination has been made, it is no longer possible to hold a House election using the previous number of seats.
In the case of a House election called prior to the completion of the redistribution process, a mini-redistribution is required. If a state loses an electorate, the two contiguous electorates with the smallest number of enrolled voters is merged. If a state gains an electorate, the two contiguous electorates with the greatest number of enrolled voters are split into three seats.
Such an outcome would be very much a stop-gap measure and would leave the voters of one or two communities with a far less powerful vote than others.
This is a dramatic process and is far from ideal. I wouldn’t rule out a government calling an election at a time that would cause a mini-redistribution, but it would be quite unusual.
So if we think about what might be different this year, one thing that stands out is the scale of the change.
For this post, I have settled on a new way to divide up redistributions. When a redistribution has been triggered by a change in the state’s seat entitlement, that’s a major redistribution. They tend to be more dramatic, and you can’t proceed with an election until it was resolved. Other redistributions I’m calling minor redistributions. In 2010, a redistribution in Victoria was paused during the holding of a federal election, but it wasn’t a problem because it didn’t involve a change in Victoria’s seat entitlement.
This year’s redistribution is the most dramatic since 1984 (when every state was redrawn at the same time). 100 of the 150 seats to be used at the next election are currently being redrawn in major redistributions (plus 2 NT seats undergoing a minor redistribution).
The number of seats being redrawn has varied significantly over the last quarter century. At the last three elections, about a third of the House was effected by major redistributions. No major redistributions were conducted prior to the 2013 election, while about half of all seats were affected by major redistributions in 2007 and 2010.
So before we consider the issue of timing, we should acknowledge that there are an unusually large number of seats where we still don’t know what the boundaries will be for the next election, despite many preselections having already been concluded.
Next up, I wanted to understand how long redistribution processes have been taking, and how that interacts with the date of the election, and the earliest possible date for a standard election (which is always in early August one year before the Senate term ends). For this chart, I’m only looking at major redistributions.
There is some variation in how long it takes before the redistribution process commences, but it’s usually at least 400 days after the previous election. The only exception was in 1996, when it took 366 days to start the redistribution. It took longer following the 2001 and 2007 elections, which were both held in November. In both years the parliament didn’t sit until the following year, which added time before the start of the process.
For the 2001-2016 elections, the previous election had been held in the second half of the year, and thus there was a shorter window between the first and last potential dates for a standard election, and thus there was a significant gap (usually around half a year) between the finalisation of the redistribution and the first possible date for an election.
The 2015-16 redistributions took longer, and that year’s election was a double dissolution was called for a date about a month before the earliest date for a standard election.
In 2019 and 2022, thanks to the reset of the 2016 DD, the earliest possible date for a standard election was quite a long time before the three years runs out, and thus the redistributions were only finalised very close to the earliest date for a half-Senate election, around July-August two years after the previous election. But as long as the election was held in May, this wasn’t an issue.
That dynamic should also play out again in 2025, but with the redistribution taking substantially longer, the time between redistribution finishing and the election is quite a lot less. The AEC’s timetables now say that the redistributions will be finalised on October 24, compared to August 2 in 2021. That’s about 10 less weeks between redistribution and election, assuming the election is held again in mid-May.
So the 2024 process is taking longer than most other years, but it is similar to 2016. Now I wanted to understand what is taking so much time. This next chart dates back to 2005, and shows how much of the time was spent in different parts of the process – public submissions, time for deliberation and map-drawing, and administrative procedures at the start and end of the process.
The AEC’s website currently says that the draft boundaries for New South Wales, Victoria and Western Australia are due in late May or early June. They always publish on Fridays, so the chart is based on all three being released this Friday, May 24. In reality it’s unlikely they’d all be released on the same day.
The length of time for consultation is very consistent, about six weeks for the first two rounds of consultation, and another six weeks after the boundaries are released.
The amount of time spent deciding on the final boundaries after the consultation has also been relatively consistent, although it was longer in 2021 than in the past.
There is always a gap after the announcement of the final boundaries before they are formally implemented, but this gap has been getting smaller. It was over two months in 2006 and 2009 but it was closer to a month in 2021. This period is relevant to the formal implementation of the boundaries and would limit the ability to call an election, but during this time it will be possible for candidates, parties, media and analysts to use the final boundaries in the meantime.
The 2015 processes took a lot longer because the process commenced in December 2014, and the consultation process didn’t commence until March of the following year. That’s not the same as this year. Instead, this year’s redistribution is taking longer because the committee is taking longer to decide on the draft boundaries.
So why might that be? I don’t want to assume this is a failure on the part of the AEC, they may have good reasons.
One likely issue is that the original population projections provided to the AEC by the Australian Bureau of Statistics (ABS) for Victoria and Western Australia turned out to be wrong, and new data needed to be provided. The AEC would have needed to start over the map-drawing process, and are also less able to rely on submissions that were based on the false information.
It’s also possible that the scale of the task is slowing things down, with two thirds of seats around the country needing to be redrawn. It might be worth asking some questions at the next JSCEM inquiry into why the process takes so long.
Whatever is happening, it does seem to be the reality that this process is taking longer than previous elections. Unless the government is willing to trigger three mini-redistributions, they really don’t have the ability to call an election until late October, which would involve an election at the end of November.
If they were to do that, they would be going to an election with boundaries only just published, with very little time to prepare to campaign. It would also be a big administrative burden to be ready to go on the new boundaries.
So while an election later this year could seem plausible for other reasons, I think the long redistribution process is a very strong reason to rule out an election in 2024.