The 2025 Federal election was the fourth held since the introduction of major changes to the Senate’s electoral system. The changes were to deal with the rapid increase in nominations, their impact on ballot paper and font size, and the distortion of results produced by increasingly labyrinthine preferences deals under the now abolished Group Voting Ticket (GVT) system.
For overseas (and local) readers unfamiliar with the Senate’s electoral system, I’ll start this post with a brief background on the Senate’s electoral system and explain the changes introduced since 2016.
I will then look at the four elections since the changes and to assess how parties, candidates and voters have reacted to the new system with main focus on the 2025 result.
To do this I am analysing the Senate ballot paper data released by the Australian Electoral Commission (AEC).
There is more detail inside the post, but I’ll quickly summarise my findings –
- The 2016 changes, and a 2022 increase in party members needed to register a political party, have seen the number of parties contesting Senate elections decrease from 52 at the last election using GVTs in 2013, to 28 at the 2025 election.
- The average number of groups on Senate ballot papers rose from 7.7 at the first divided Senate ballot paper election in 1984 to 33.7 at the last GVT election in 2013 before falling to 17.2 in 2025. The result has been smaller ballot papers and a return to larger and more readable font sizes.
- At the 2025 election, 81.2% of voters completed their Senate ballot papers with a 1-6 sequence of preferences above-the-line (ATL). 5.8% completed fewer than 6 ATL preferences, 6.6% filled more than 6 ATL preferences, and 6.3% of ballot papers counted as ‘below-the-line’ BTL votes. There has been little change over four elections in the percentage of ballot papers in each completion category.
- Fewer groups and smaller ballot papers are likely to have played a part in the decline in the number of ballot papers with “exhausted” preferences at the end of the count.
Abbreviations used in this post
- ATL – ‘Above the line’, a vote cast for a party or parties above the thick black horizontal line on the Senate ballot paper
- BTL – ‘Below the Line’, a vote cast for candidates below the black line
- GVT – ‘Group Voting Ticket’,
A further definition. An “exhausted” vote is one that, when examined for transfer in a distribution of preferences, has no further preferences for candidates remaining in the count.
A Brief Background on the Senate’s Electoral System
Proportional Representation by Single Transferrable Vote (PR-STV) has been used to elect the Senate since 1949. Senators are elected by achieving a quota of votes, currently 14.3% or one-seventh of the vote for six state vacancies at half Senate elections. A candidate’s quota can be based on first preference (number 1) votes for the candidate, on preference transfers from excluded candidates, and surplus to quota preferences transfers from elected candidates.
A major change to the system in 1984 divided the ballot paper with a thick horizontal line (See example below) and introduced two voting options. Electors could vote for parties ‘above the line’ (ATL) or for candidates ‘below the line’ (BTL). By different methods before and after the 2016 changes, an ATL vote was imputed to have a sequence of preferences for candidates.
From 1984 to 2013, electors could vote for only a single party ATL, or were required to number preferences for 90% of candidates below the line. The vote was imputed to be for the party’s list of candidates as well as the party’s full list of preferences for all other parties and candidates. These preferences were registered by a ‘group voting ticket’ (GVT). Voters wanting to choose between candidates or select their own between-party preferences were required to use the BTL option and number 90% of candidate boxes for a formal vote.
The ease of ATL voting compared to BTL resulted in around 95% of voters using the option between 1984 and 2013. This put more than 90% of distributed ballot papers votes under the control of GVTs. Even parties with low levels of support and miniscule campaigns could deliver tight preference flows using GVTs. Preferences became tradeable commodities, and over three decades, instances grew of parties with very low vote shares defeating higher polling competitors via labyrinthine preference deals.
Reforms of the Senate’s electoral system in 2016 abolished GVTs and ended the ability of parties to send preferences to other parties. Since 2016 only voters can send a vote to another party by preferences.
A new form of ATL voting was introduced allowing voters to complete an ATL preference order for parties. Ballot paper instructions state that a voter should complete six preferences for an ATL vote or 12 for a BTL vote. A savings provision for ATL votes permit any vote with at least an ATL first preference to remain in the count as formal.
The old system had led to a proliferation of parties contesting election and swapping preferences via GVTs as a tactic to accumulate or ‘harvest’ preferences. The tactic was ended with the abolition of GVTs.
Registration of parties was also changed in 2022 by the number of members required for registration increasing from 500 to 1,500.
The combined impact of these changes has been to reduce the number of parties contesting election and the number of columns on Senate ballot papers. The result has been smaller ballot papers and a return to larger and more readable font sizes.
- The first election using the divided Senate ballot paper in 1984 was contested by an average 7.7 groups per state.
- The last GVT half-Senate election in 2013 was contested by 52 parties with a quadrupling of the average columns per state of 33.7.
- The number of parties contesting the Senate election dipped to 48 in 2019, before the higher membership requirement reduced numbers further to 40 in 2022 and 28 in 2025.
- The average number of columns per ballot paper has almost halved since the end of GVTs in 2013, falling to an average 17.2 columns per state in 2025.
Ballot Paper Completion Categories by State
The table below breaks down ballot paper completion rates based on ATL versus BTL, and further divides ATL votes into categories based on the number of preferences completed. 2025 figures are broken down by state and territory with 2016, 2019 and 2022 national totals included for comparison.
The table shows very little change across the four elections. Around 81% of voters follow the ballot paper instructions and number preferences 1-6 without completing further preferences. Around 6% remain formal by the savings provision having completed fewer than six valid preferences. Around 8% go beyond six preferences and around 5% are ballot papers completed below the line. The table shows little evidence of change in completion rates over four elections.
There are differences by state. NSW always has more ballot papers that stop before six ATL preferences. Familiarity with the Hare-Clark electoral system sees more Tasmanian and ACT voters vote below the line. Smaller ballot papers in Tasmania and the two territories see more people go beyond 6 ATL preferences.
| Number of ATL Preference | ||||||
|---|---|---|---|---|---|---|
| Jurisdiction (Groups) | 1 | 2-5 | 6 | 7-12 | >12 | BTL |
| New South Wales (18) | 4.6 | 3.8 | 81.3 | 4.3 | 1.8 | 4.2 |
| Victoria (20) | 1.9 | 2.9 | 83.4 | 5.2 | 1.8 | 5.0 |
| Queensland (19) | 1.4 | 2.7 | 84.1 | 4.8 | 2.2 | 4.8 |
| Western Australia (18) | 2.2 | 2.7 | 83.2 | 5.1 | 3.0 | 3.7 |
| South Australia (16) | 2.2 | 3.2 | 80.1 | 5.5 | 3.2 | 5.8 |
| Tasmania (12) | 1.1 | 2.1 | 63.5 | 15.0 | .. | 18.4 |
| Australian Capital Territory (7) | 1.2 | 1.7 | 47.2 | 33.2 | .. | 16.8 |
| Northern Territory (8) | 1.9 | 2.8 | 49.3 | 36.3 | .. | 9.8 |
| 2025 National Total | 2.7 | 3.1 | 81.2 | 5.8 | 2.0 | 5.2 |
| 2016 Election | 2.6 | 3.5 | 81.6 | 4.9 | 0.8 | 6.5 |
| 2019 Election | 3.5 | 3.5 | 80.0 | 4.5 | 1.2 | 8.3 |
| 2022 Election | 2.6 | 3.2 | 81.3 | 5.3 | 1.3 | 6.3 |
Ballot Paper Completion Categories by Party
The table below shows completion categories by party. The patterns by party are similar to those seen at previous election.
The two major parties, Labor and the Coalition, are the only two where 1-6 ATL voting rates are above 80%. Both have lower rates of voters going beyond 6 ATL preferences or taking the option to vote BTL.
The lower the recorded votes for a party, the lower the rate of voters stopping after 6 ATL preferences, which as a consequence sees higher rates of voters going beyond 6 ATL preferences, and higher rates voting below the line.
For all parties that contested the Tasmanian Senate, there were far higher rates of voters going beyond 6 ATL preferences or voting below the line. For the Jacqui Lambie Network in Tasmania, only 52.2% of votes were completed 1-6 ATL, and 31.6% voted below the line, the highest rate in the nation.
| Number of ATL Preference | ||||||
|---|---|---|---|---|---|---|
| Party (% Vote) | 1 | 2-5 | 6 | 7-12 | >12 | BTL |
| Labor (35.11) | 3.0 | 2.8 | 83.5 | 5.1 | 1.7 | 3.9 |
| Coalition (29.89) | 2.8 | 3.3 | 85.3 | 4.1 | 1.4 | 3.2 |
| Greens (11.72) | 1.1 | 2.2 | 78.4 | 7.4 | 3.0 | 8.0 |
| One Nation (5.67) | 2.4 | 3.5 | 78.5 | 7.1 | 2.9 | 5.5 |
| Legalise Cannabis (3.49) | 4.1 | 4.8 | 75.3 | 7.0 | 2.6 | 6.2 |
| Trumpet of Patriots (2.60) | 2.7 | 3.5 | 79.6 | 6.5 | 2.8 | 4.9 |
| Family First (1.49) | 2.8 | 4.0 | 76.9 | 6.1 | 2.4 | 7.8 |
| Rennick People First groups (1.27) | 0.9 | 2.4 | 74.3 | 9.0 | 3.4 | 10.0 |
| Animal Justice (1.25) | 2.8 | 4.1 | 76.4 | 6.6 | 2.6 | 6.0 |
| Jacqui Lambie Network (1.05) | 1.9 | 2.7 | 75.3 | 6.4 | 2.2 | 11.4 |
| Australia’s Voice (0.75) | 3.6 | 4.6 | 76.3 | 6.4 | 2.8 | 6.4 |
| Australian Christians (0.65) | 3.6 | 3.9 | 77.3 | 6.6 | 2.6 | 6.0 |
| Socialist groups (0.64) | 0.8 | 2.3 | 59.3 | 12.4 | 5.9 | 18.6 |
| Indigenous Party (0.64) | 2.9 | 4.7 | 72.9 | 7.3 | 2.7 | 9.5 |
| Libertarian (0.40) | 3.3 | 3.8 | 70.5 | 9.6 | 3.5 | 9.3 |
| Sustainable Australia (0.37) | 2.4 | 3.2 | 64.5 | 13.7 | 4.3 | 11.9 |
| Fusion (0.29) | 2.5 | 3.7 | 65.7 | 11.0 | 5.4 | 11.7 |
| Citizens Party (0.22) | 4.7 | 4.0 | 69.5 | 8.5 | 3.3 | 10.1 |
Rates of Ballot Paper Exhaustion and the Causes
The chart below plots final count exhaustion rates for the six states at elections since the new Senate system was introduced in 2016. The overall rate of exhaustion
I suspect that the smaller number of parties on ballot papers is playing a part in the decline in exhausted preferences at the end of the count. On the right of politics, One Nation has emerged as a magnet for preferences from other right wing parties, and by being present at the end of the count in all five mainland states, it resulted in fewer votes exhausting.
Exhaustion rates are always lower where the final race resolves into a left versus right contest. Four of the five mainland states resolved into a contest between Labor and One Nation for the final seat creating a clearer left-right contest and lower exhaustion rates, between 2.6% in SA and 4.5% in Victoria.
In Queensland, where the final contest was between One Nation and Gerard Rennick’s People First, two right-wing parties, the rate of exhaustion was a higher 5.8%. Understandably, 81.8% of Green voters, 70.5% of Socialist Alliance voters and 67.4% of Labor voters made no choice between One Nation and Rennick.
Commonwealth of Australia Senate elections no longer have compulsory preferential voting. This has led to a highly desirable decline in informal votes in the Senate.
However in the House of Representatives we maintain compulsory preferential voting. Evidence suggests that the larger the number of candidates, the higher the informal vote, disproportionately at heavy Labor voting booth. If we are happy to limit compulsory preferential voting to six in the senate why shouldn’t we do the same in the House of Representatives?
The decline in informal votes in the Senate has a direct correlation with a reduction of the number of parties contesting since the abolition of GVTs. I suspect this decline will continue as minor parties conclude that their is little value contesting elections that they cannot win.
I assert that, although every informal vote is bad for our democracy, so too are exhausted votes undesireable. Unlike Roger, instead of removing compulsory preferential voting from the House, I would favour introducing it in the Senate ATL.
For voters first preferencing a minor party, they run the risk of their vote not discriminating between lesser preferred parties that may be in the contest for the final seat, unless they complete all preferences.
The reduction in parties contesting will reduce the informal vote, for sure, and with the advent of electronic voting the computer can assure that a vote is formal before submission.
COMMENT: I doubt that Senate groups are going to go below 10-15. Tasmania was down to 12 groups in 2025. The average informal voting rate when full preferential voting applied between 1934 and 1983 was 10%. And if you have full preferences above the line, will you require it below the line? If not, are optional preference instructions below going to conflict with full preferences instructions for above the line. I don’t think any participants are interested in pushing for ATL full preferential voting given it would increase the informal vote.