Skip to the content.

How Senators are elected



Just as with the House of Representatives, Australia uses preferential voting to elect Senators, but our votes are counted using a different mechanism that gets us proportional representation. In this blog post, we’re going to learn how to fill out a Senate ballot, and how our Senate ballots are counted.

This article is mostly independent of political opinion. Where opinion creeps in, it will be explicitly flagged as such, like below:

Opinion: Kangaroo is tastier than beef.

If you haven’t read How voting for your House of Representatives Member works, then please give that a read first: we’ll be making references to that often.

Before making this article publicly available, I asked Kevin Bonham, a psephologist, to review it, and provide commentary on any inaccuracies so that they might be fixed. Although they haven’t officially approved the content, with their input, I am confident that it’s an accurate representation of our voting system. However, you should trust all communication by the Australian Electoral Commission (AEC) above what you read here.

In this post, we’ll discuss:

  1. Why we can’t use the same counting method in both the House and Senate
  2. Understanding the Senate ballot
  3. Case Study
  4. Who really does the preferencing (Senate edition)?

Why we can’t use the same counting method in both the House and Senate

When electing someone to the House of Representatives, there are multiple candidates running for one seat that represents somewhere between 70 000 and 140 000 voters (AEC enrolment stats). There is always exactly one winner, which means that we’re able to find someone who accumulates more than half the votes through successive rounds of elimination.

Unlike in the House, there are twelve Senators for each of the six states and two each for the ACT and NT. In a normal election, we elect about half the Senators, so six for the states and two for the constitutionally-recognised territories. When there’s a double dissolution, we need to elect a whole new Senate. In this situation, the states need to have twelve winners, and the two territories still have two winners. If we use the same method for the House as we do in the Senate, that means that over half of the state would agree on one Senator, but what about the others?

To get around this problem, we employ a counting method called single-transferable vote for our Senators. It’s still a form of preferential voting, but it enables us to determine multiple winners instead of just one. This enables each State and Territory to have proportional representation, which means that your vote contributes to electing all senators, not just one.

Formal ballots and informal ballots (again)

As with the House of Representatives ballot, provided you fill out your ballot paper according to the rules, it’s impossible for you to waste your vote. Here are the main rules again (in list format):

We call a ballot paper that’s followed all the rules a formal ballot. Only formal votes are counted and contribute to the election result. The AEC provides a lengthy guide outlining what constitutes formal and informal ballots and they also have a practice mode.

Understanding the Senate ballot

Sample practice ballot provided by the Australian Electoral Commission.

The Senate ballot paper is white and looks like the sample ballot above. The thick black line divides two separate ways of voting: we can either number the boxes above it, or we can number the boxes below it: but we can’t do both. There are also instructions to the left that tell you the minimum number of boxes you need to number for your method.

When we vote Below the Line (BTL), we are voting for individual candidates, just like in the House of Representatives. We can number the boxes however we want.

Voting Below the Line

Example filling out for Below the Line, with votes randomly assigned to fourteen of seventeen candidates.

In this example, we can see that the voter has randomly numbered the boxes below the line, and has even left some of the boxes blank. Because the instructions said to number “at least 12 boxes”, and the voter wrote in 14, their ballot is still formal.

Voting below the line is the only way to vote for or give preferences to an UNGROUPED candidate.

Voting Above the Line

Example filling out for Above the Line, with votes randomly assigned to six of seven groups.

For convenience, we can also preference groups of candidates. Voting above the line tells the counters to treat our ballot as if we had numbered the boxes below the line in ascending order for the group. That means that the ballot just above this paragraph is exactly the same as the following one!

Example filling out for Below the Line that's identical to our Above the Line example.

This kind of voting is relatively new: it only came about in 2016. There was a different method for Above the Line voting prior to the 2016 federal election. If you’re a voter older than 24 and aren’t an election nerd like me, then this might come as a surprise. We’ll talk more about this in the preferencing section at the bottom.

Remember: there’s no way to vote for or give preferences to an UNGROUPED candidate when you vote Above the Line.

Case Study

This section has a lot of numbers in tables. If you don’t like numbers, you can skip down to the preferencing section, so long as you’re still doing preferential voting.

We’re not going to use a real-world scenario in this case study due to the sheer length of the process. In the 2019 federal election, here’s how many rounds were necessary for each state:

State Senators to elect Rounds
NT 2 1
ACT 2 25
TAS 6 139
SA 6 177
WA 6 231
QLD 6 288
VIC 6 367
NSW 6 429

The NT election isn’t interesting because it happens too quickly, and the ACT has far too many rounds in what’s already, at this point, a long article. Fortunately, the AEC has a nice infographic that covers the process for us with fewer rounds than the ACT, so we’ll defer to that for explaining the counting mechanism. We spend the rest of this section discussing how things work and why. Everything in this section is derived from the AEC’s official page, ‘How the Senate result is determined’.

Determining the minimum number of votes to get elected

The first thing we need to do is establish the minimum number of votes a candidate needs in order to be elected. When there’s only one vacancy, we just need more than half the total number of votes. When there’s multiple vacancies, we apply the formula (number_of_voters / (vacancies + 1)) + 1. For the 2019 election, the quotas looked like this for each state:

State Senators to elect Voters Formula Quota
NT 2 105’027 (105'027 / 3) + 1 35’010
ACT 2 270’231 (270'231 / 3) + 1 90’078
TAS 6 351’988 (351'988) / 7 + 1 50’285
SA 6 1’094’823 (1'094'823 / 7) + 1 156’404
WA 6 1’446’623 (1'446'623 / 7) + 1 206’661
QLD 6 2’901’464 (2'901'464 / 7) + 1 414’495
VIC 6 3’739’443 (3'739'443 / 7) + 1 534’207
NSW 6 4’695’326 (4'695'326 / 7) + 1 670’761

You might be wondering if it’s fair that Senators running for the NT only need 35 010 votes, while NSW needs a whopping 670 761. Remember that we’re aiming for proportional representation within each state and territory. That not only means we want to have all voices heard for all vacancies, but also that the number of votes required to get elected needs to increase with the number of voters in the electorate.

The counting process

Step 1: This is still preferential voting

Since this is still a form of preferential voting, we first allocate the votes into groups corresponding to voters’ first preferences. Any candidates who meet the quota are immediately elected. The 2019 election saw at least one candidate immediately elected for each state (two in the NT’s case, which is why there was only one round).

Step 2: Transferring the surplus

Popular candidates, such as independents who are extraordinarily active within their community and candidates for the louder parties, will probably end up with more votes than the necessary quota. As we discussed above, that’s not a good situation to be in when there are multiple candidates. To ensure that proportional representation is achieved, if there’s even a single vote over the quota for a candidate, then we need to transfer that surplus to the second preference. There’s a problem though: whose ballots are considered to be a part of the “surplus”? I consider it natural that almost everyone would want their vote to be included in the surplus that’s transferred, but if we transferred everyone’s votes, then we’re moving a lot more than the surplus!

Fortunately, there’s a way to achieve this without moving more than the surplus’ amount of votes. Everyone’s votes are transferred, but with a “voting power” to match the surplus. The formula applied here is voting_power = surplus / total_votes_for_candidate. For example, if Candidate A gets 1’000’000 votes, but the quota was only 500’000, that means that everyone who Voted 1 for A will have their 2 Vote considered at half “strength”.

Once we’ve moved everyone’s votes over at the reduced “strength”, we check to see if anyone has met or exceeded the quota, and repeat if necessary. Votes with a reduced “strength” that contribute to another surplus will have the same “strength” as a vote that’s being transferred for the first time. For example, if Candidate B had enough votes transferred so that they ended up with 750’000 votes, and some of those votes were from the half-“strength” Candidate A transfer, then those votes will have ⅓ “strength” for their 3 Vote, just like the people who Voted 1 for B’s 2 Vote. The fairer method would be for those 3 Votes to continue to weaken, meaning that they’d have ⅙ “strength” for the 3 Vote and so on, but successive governments just haven’t fixed this problem. Antony Green wrote about this in 2014.

This process is repeated until there are no more surpluses, or until all the seats are filled.

Step 3: Candidate elimination

Once we end up in a situation where there are no more candidates who have reached the quota, we go back to the rules of single-seat preferential voting. That means that the unelected candidate with the fewest votes is eliminated, and the votes allocated to them are redistributed to the voters’ second, third, etc., preference (depending on how far they travelled in the suprlus stage). A key difference between Step 2 and Step 3 is that votes moved in Step 3 move with their current “voting power”. For example, if Candidate C is eliminated, then everyone who Voted 1 for C will have their votes moved at full “strength”, while people who Voted 1 for A and 2 for C will have their votes moved at half “strength”.

We then jump back to Step 1 to repeat the process, until there are no more vacancies.

Corner case 1: What happens when counting gets to an empyty box?

Remember that we don’t need to fill out all the boxes to have a formal Senate ballot: we just need to fill in at least the minimum number of boxes for our voting method. That means many ballots are probably going to have empty boxes after a while. Once your vote gets to that point, it means that you’ve told the counters you no longer care who is elected, and your ballot will be set aside and no longer participates in the process.

Corner case 2: What happens when there’s still one or more vacancies and not enough ballots for anyone to meet the quota?

In this case, we still continue redistributing the ballots until we’re left with just two candidates. The candidate with the majority of votes will be elected.

Such a complex process! Why do we have it?

Voting for Senators is as easy as voting for your local representative. It might be a bit lengthier, but the rules remain exactly the same for voters. Counting, on the other hand, is more complicated, but it ensures that it’s impossible to waste your vote, and that your voice is heard as much as possible.

Opinion: The complexity of this method actually makes it one of the fairest voting systems in the world.

Who really does the preferencing? (Senate edition)

The rules for Senate voting changed at some point in 2016, prior to the federal election that year. That means that if you’re over the age of 24, you might be familiar with the concept of Group Voting Tickets. Group Voting Tickets have been abolished at the federal level, and also at the state level for NSW, SA, and WA. That means in the upcoming federal election, there are no preference deals among parties and parties without sufficient votes to meet a quota cannot give your vote away.

Parties can provide How-to-Vote cards, which is about the closest thing we’ve got to “preference deals”, but it’s merely a recommendation on their part. For example, if the sample ballot above had run in an election, there would have been a maximum of 8 How-to-Vote cards: one for each group above the line, and one for each ungrouped candidate. There are over 966 trillion different ways to cast a formal vote on that ballot! If you don’t want to follow a How-to-Vote card, you can simply refuse them or put them in the bin.

Just as when voting for your local representative, only you decide your vote. Anyone telling you otherwise is spreading a myth from a bygone era of Australian history.


In this blog, we saw that voters get to vote in the same way on both ballots (but don’t need to number all of the boxes), that the counting process is very involved (but also very fair), and that only you decide your vote. Please share this with your friends and family to ensure that they understand our wonderful voting system.

If the article is popular enough, I’ll write a third one to talk about how to maximise your vote.