On election night, the AEC is legally required to conduct an indicative count in each House of Representatives contest. After first preferences votes are allocated and counted votes are re-sorted into two piles – these piles are for the candidates deemed by the AEC as most likely to be the final two candidates in the count.

Why does this happen?

Believe it or not, this is about television coverage. Well, that and the ability to have as early an indication as possible.

In the early 1990s, there was some dissatisfaction on the part of television viewers (as well as television producers!) that the full preferential counts the AEC conducts take too long, meaning that it was difficult for election analysts to make predictions about winners and losers on the night.

As a result, electoral laws were amended to require these early, indicative counts.

Who decides which candidates to use for these counts?

Before election day, the AEC makes a decision about which two candidates are most likely to be the final two candidates in each seat after preferences are distributed. These decisions are based on factors like previous results, media coverage, opinion polling and impressions on the ground in each electorate.

As you might expect, the identification of the two candidates who will be part of that initial indicative count for each seat is not released until after polling is complete at 6pm local time.

What if you get the top two candidates wrong?

This happens in some seats and it’s OK. We simply stop the initial two-candidate count and start a fresh one with those who are truly the top two candidates in the seat. We always follow what the votes are telling us, and the result will always reflect what is on ballot papers.

At the 2022 federal election, the AEC stopped and re-ran two-candidate preferred counts in about a dozen seats.

Do you still distribute all the preferences in every seat?

Absolutely. The AEC runs full preference distributions for every House of Representatives seat – even those in which a candidate gets more than 50% of primary votes and wins without needing preferences to be distributed.

Two-candidate preferred counts give election analysts, voters and candidates useful information about the likely result in each seat, without making them wait. Preference distribution is a complicated process, however, and it can’t be done until every possible vote has been received by the AEC and added to the count. Because the AEC needs to leave 13 days after election day for postal votes to get back to us, this takes a lot of time.

In a small number of House of Representatives seats, it can be difficult to determine which candidates to select for the initial two-candidate preferred count. This can happen in seats where it is a tight three-way contest or in other seats where it is just difficult to predict who will be second or third.

So what do we do?

In these cases, the AEC may conduct a three-candidate preferred count instead – sorting votes into piles for three candidates instead of two. Like the two-candidate preferred count, ballot papers go into the pile based on who received the highest preference. In practice, this means we set aside ballot papers with 1st preferences for the 3 leading candidates, then we take ballot papers with 1st preferences for the remaining candidates and allocate each ballot paper to one of the 3 leading candidates based on the 2nd preference on that ballot paper (or the next preference available for one of the 3 candidates). Then we tally up how many votes each of the 3 candidates have to understand who the likely winner will be, and who the top two candidates will be.

Once the two leading candidates are determined the legislated two-candidate preferred count will be conducted, which as always is followed by the full Distribution of Preferences

When has this happened?

At the 2022 federal election three-candidate preferred counts were undertaken in the seats of Brisbane (Queensland) and Macnamara (Victoria). There have been three-candidate preferred counts undertaken in other recent federal elections, but it has not been a common requirement.

In Brisbane, one candidate received 37 per cent of the primary votes, however the next two candidates received an almost identical number of primary votes – their totals only differed by 11 votes. A three-candidate preferred count was required to determine whether the Labor or Greens candidate would be excluded from the count first, with that candidate’s votes then being distributed into a traditional two-candidate preferred count.

In Macnamara, the distribution of first preference votes was very close between the Liberal, Labor and Greens candidates. A three-candidate preferred count was required in order to determine which two of the three leading candidates to use in the two-candidate preferred count.

How does the AEC communicate three-candidate preferred counts?

The AEC’s Tally Room was not designed with three-candidate preferred counts in mind. For Brisbane and Macnamara in 2022 the results of three-candidate preferred counts were published to webpages on the AEC’s website. At the conclusion of each three-candidate preferred count, the Tally Room was then used to publish the results of a standard two-candidate preferred count.

This practice will be repeated at the 2025 federal election while the AEC works on upgrades to the Tally Room that will allow it to track three-candidate preferred counts at future elections.

To win an election outright and be invited by the Governor-General to form government, a political party (or a coalition of parties) needs to win more than half of the 150 seats in the House of Representatives – that’s at least 76 seats.

Historically, Australian federal elections have often produced majority governments - since the turn of the century, one party or coalition has won at least 76 seats in all but one election.

But what happens when no party or coalition wins 76 seats?

Time to negotiate

The AEC doesn’t have a role in negotiations around minority governments - other than to strive to have results in each electoral division known as efficiently as possible to facilitate negotiations.

If a party or coalition doesn’t have 76 seats in its own right, attention typically turns to what’s called the crossbench. These are the seats in the House of Representatives that sit in the middle, between the Government on one side of the Chamber and the Opposition on the other. Crossbench members are likely to be members of minor political parties, or independent of any political party.

For example, if a party or coalition wins 73 seats in the House of Representatives, and is able to convince three or more crossbench Members to guarantee their support, they can then be invited to form a Minority Government by the Governor-General. The crossbench Members don’t necessarily have to join the party (or parties) of the Government, or even form a formal coalition. The important thing is for the Government to be able to demonstrate that it commands the support of the majority of the Members in the House of Representatives.

Negotiations about forming a Government where no party or coalition commands a majority of seats typically happen in the days after a federal election, once it becomes clear that no majority will be achieved.  

Why 76 seats?

The size of the House of Representatives for the next federal election is 150 seats. Half of 150, plus one, equals 76. Simple maths tells us that if a party, coalition or group is able to command the support of 76 seats, the opposing party, coalition or group will only be made up of a maximum of 74 seats – not enough to win a vote in the House of Representatives, and therefore not enough to form a Government.

Why is it called minority government?

It’s an informal term, and often used to differentiate this type of government from a majority government, which is one in which the governing party or coalition holds more than 76 seats.

When has this happened?

There have been two recent examples of minority government in Australia.

In 2010, the Australian Labor Party won 72 seats, while the Coalition of the Liberal Party, the Liberal National Party of Queensland, the Nationals and the Country Liberals won a total of 73 seats. The remaining five seats were held by four Independent Members and one Member representing the Greens.

In this case, the Labor Party led by Prime Minister Julia Gillard ultimately reached an agreement with the Greens and three of the four Independent Members, for a total of 76 seats. The Coalition was therefore only capable of reaching 74 seats and formed the Opposition. Specific detail about this is available via a 2010 Parliamentary Library paper.

At the 2016 federal election, the Coalition was re-elected with a total of 76 seats in the House of Representatives and formed a majority Government. Subsequent to this, however, one of the seats held by the Government was won by an Independent at a by-election, and two Members of the Government moved to the crossbench. By the time the Parliament was dissolved for the 2019 federal election, the Government held 73 seats and was technically in minority before once again winning a majority of seats at that election.

At every federal election, there’s a good chance that a small number of House of Representatives seats will be decided on a close margin, with the difference between the winning candidate and the runner-up coming down to a matter of a few hundred votes, or sometimes even less.

Regardless of whether a result is close or wide, the AEC’s first priority is always the integrity of the count.

When has this happened?

Close margins happen in every federal election.

At the 2022 federal election, the seats of Gilmore (New South Wales) and Deakin (Victoria) were decided by margins of 373 and 375 votes, respectively.

At the 2016 federal election, the seat of Herbert (Queensland) was decided by a margin of just 37 votes. At the 2013 federal election the seat of Fairfax (Queensland) was decided by a margin of 53 votes. These margins were so close that a full recount was undertaken in each seat to ensure that everybody involved had confidence that the result was correct.

What is the practical effect?

A close count simply means it will likely take some time before the result in that seat is known. This is because it may require every vote cast for the seat to be returned to the AEC, counted and a full preferential count of all ballot papers to be undertaken. This is as opposed to other seats with larger margins where people are able to get a firm indication of who will win prior to these activities being finalised.

For example, if a seat has a margin between the two leading candidates of 500 votes and our systems tell us that there are 500 or more postal voters yet to return their postal votes for that seat, we cannot declare that result until the number of potential votes left to be received is less than the margin.

For postal votes and votes cast interstate or overseas it takes 13 days after election day (the legislated cut-off for receipt of postal votes) for all votes to be available for the full distribution of preferences to occur.

The Australian Electoral Commission counts every ballot paper at least twice. The initial count of ballot papers starts on election night and continues as postal, interstate and overseas votes are received in the days afterwards. A mandatory second count of every vote - something called ‘fresh scrutiny’ - is undertaken to double check the initial count. This occurs in the period after election day as soon as practicable. It is not uncommon for small changes in count totals to be seen when this secondary count is undertaken – this is typically the result of administrative errors where totals have been tabulated. Picking up and correcting these issues is precisely what the secondary count is for.    

When the margin in a seat is particularly close, it’s not unusual for candidates or political parties to request a recount. If accepted, this can be a full recount where the AEC goes back to square one and counts every ballot paper again. It can also be a partial recount relating to a particular type of votes (e.g. declaration votes cast outside the division).

Does the AEC have to run a re-count?

At any point before the AEC officially declares the result in a seat, a candidate can make a request to that Division’s Returning Officer for a full recount of every ballot paper. The candidate is required to provide reasons for the request, and it’s the responsibility of the Returning Officer to approve or decline that request.

If a Returning Officer declines a request for a recount, candidates sometimes appeal the decision to the Electoral Commissioner.

The AEC can also conduct a recount of votes independently. While there is no law for when this activity occurs, the AEC often conducts a recount for a seat if the final margin is within a hundred votes.

What happens during a recount?

If a recount is agreed to, the Divisional Returning Officer is required to notify each candidate in the electorate that a recount will occur, and the time and place it will occur.

The process of a recount is largely the same as an initial count of the ballot papers, with the exception that there is no need to wait for postal or declaration votes to be returned as they are already on hand. All preferences are checked to ensure they have been appropriately allocated – meaning it can take some time. Scrutineers appointed by candidates can observe a recount, just like an ordinary election count.

When a recount is complete, the Divisional Returning Officer will make a formal announcement of the result, known as the declaration of the poll. These are public events.

When has this happened?

At the 2022 federal election, a request was made for a recount in the seat of Gilmore (New South Wales), which was won by a margin of 373 votes. The request was declined on the basis that there was insufficient grounds for the activity to occur.

At the 2016 federal election, the Divisional Returning Officer for the seat of Herbert (Queensland) made the independent decision to conduct a recount for the seat due to the margin being under 100 votes. Following the recount and a full distribution of preferences, the final margin in the seat of Herbert was found to be 37 votes.

The decision on when to invite a party or coalition to form Government, and when that Government can be officially sworn in, is a responsibility of the Governor-General, and not the AEC.

Just like all Australians, the Governor-General may rely on the AEC’s early, indicative counts for insight into who has won the election, and which party, coalition or group might be in a position to form Australia’s next Government.

Why is this important?

When an election is called, the House of Representatives is dissolved and Australia’s Government enters what’s called caretaker period.

During the caretaker period, the business of government continues and ordinary matters of administration still need to be addressed. However, successive governments have followed a series of practices - the ‘caretaker conventions’ - which aim to ensure that their actions do not bind an incoming government and limit its freedom of action.

The swearing in of a new Government (be it from the same political party as the previous one, or not) allows the full business of Government to recommence.

Can it happen quickly?

Yes, it did happen swiftly for the 2022 federal election.

At the 2022 federal election, the Coalition Government was defeated by Labor Party. While counts for some seats remained close, it became clear on election night (Saturday 21 May 2022) that the Coalition would be unable to secure 76 seats in the House of Representatives. In a televised speech, the previous Prime Minister conceded that his Government had been defeated and congratulated the Labor Party on their victory in the election.

On Monday 23 May 2022, Mr Albanese was sworn in as Australia’s 31^st Prime Minister. While not as quick as 2022, for the 2019 federal election Mr Morrison was also sworn in as Prime Minister prior to the return of writs.

The AEC is not a party to any of these events. Our role is to count votes and provide early indicative data about which candidate is likely to win in each seat. The formation of Government and the ability to command a majority of Members of the House of Representatives is a matter between political parties and the Governor-General.

Are there examples of it being slower?

Yes.

Following a close result at the 2016 federal election, Prime Minister Malcolm Turnbull was not sworn in until 16 days after election day. With the result so close, Opposition Leader Bill Shorten did not concede defeat in the election until eight days after election day.

At the 2010 federal election, held on Saturday 21 August, Prime Minister Julia Gillard was sworn in on September 14, nearly a month after election day. The delay in this case was caused by a need to determine the final makeup of the House of Representatives, and following that for both the Labor Party and the Coalition to negotiate with crossbench Members in order to form a majority in the House of Representatives.

When you number your ballot paper, the ‘1’ you write in the box for your most preferred candidate is called a primary vote, or simply your first preference.

First preferences are an important part of counting the vote – in fact, a number of seats around the country are often decided on first preferences, with one candidate winning 50% + 1 of the vote and no need to allocate voter preferences to determine the winner. However, the AEC still conducts a full distribution of preferences in every seat, regardless of whether one candidate wins 50% + 1 of the vote. For a growing number of seats, first preferences only tell us part of the story.

How can a party or coalition form government if it doesn’t get 50% of the primary vote?

Australian elections are all about preferences.

At the 2022 federal election, only 15 seats were decided on first preferences alone. For each of the remaining 136 seats, preferences needed to be distributed in order to arrive at a winning candidate. This means that someone could give their primary vote to a smaller political party or an Independent, but that vote could ultimately flow to the winner in that seat if that wasn’t an independent or smaller party.

A federal election is a national election, but you can also think of it as 150 distinct elections happening in seats around the country. For most of these seats, preferences are needed to determine a winner.

Don’t forget, winning government in Australia isn’t about the raw number of votes your party (or coalition) gets. It’s about how many seats you win in the House of Representatives.

When has this happened?

Frequently and consistently.

Political parties often form government without receiving over 50% of the primary vote. In fact, for the largest election wins in recent history (for the curious: John Howard winning 86 seats and control of the Senate in 2001, Kevin Rudd winning government in 2007, and Tony Abbott winning government in 2013), no party received more than 46% of the primary vote.

Shouldn’t a party with 30% of the vote win 30% of the seats in the House?

Not necessarily – it depends on where those votes are cast. Remember, an election is really 150 smaller elections happening around the country. A party might get a lot of primary votes in a seat that it wins on first preferences, but relatively few for a seat that it wins on preference distribution.

Are first preference votes a useful way to measure election winners and losers?

Yes and no.

Primary votes aren’t very useful when it comes to measuring whether an election has been won or lost, because of the factors discussed above. There are even examples from recent history in which a party’s primary vote declines, but it goes on to win more seats in the House.

Perhaps the most useful number to follow is the number of seats won by a political party or group. This is the number that will ultimately determine whether a party or coalition is able to form government.

On the other hand, first preference votes play an important role. First preference counts for each candidate are some of the first reported on election night, and that data is often used by election analysts to make early predictions.

There’s also a financial component to first preference votes. After every election, as per legislation, the AEC allocates election funding to eligible candidates and parties based on the number of first preference votes they received.

Would Australian elections be different if we only used first preference votes?

Very probably yes.

The system described in the question is something called First Past the Post, and it’s used in a number of other countries. In a First Past the Post system, preferences don’t exist, so the candidate to win the most votes is simply declared the winner. Australia has not used this system since 1917.

It’s very difficult to compare Australia’s current electoral system to a hypothetical First Past the Post system, because you have to consider voter behaviour. In our current system, voters are told to number candidates in the order of their preference. Most voters know that these preferences are then distributed to produce winners in each seat. A voter might therefore choose to vote for a particular candidate first and then rank a separate candidate they think may win as a backup option of sorts, safe in the assumption that if their preferred candidate does not win their vote will at least be allocated to someone other than who they least want elected.  

In a First Past the Post system, your vote doesn’t get a ‘second chance’ if your preferred candidate doesn’t win. It’s safe to assume that if this kind of system was used in Australia, many voters would take it into consideration and potentially change their votes according to who they believe has a realistic chance of winning (or how to vote to ensure who they most dislike does not get elected).