Australia says 90.7% of enrolled voters voted in 2025. The real number is 84.1%. Almost 2.9 million eligible Australians had no say. In the Northern Territory, turnout fell below 71%. In Lingiari, it was 62%. The fine for not voting is $20. It has not changed since 1984. About 47% of Gen Z voted just to avoid the fine. It was not out of civic duty. The AEC has never surveyed non-voters about why they stayed home. Victoria did, and found 42% did not know the election was on. One in six Australians did not cast a valid vote. That is the real participation rate.
Australia has compulsory voting. The headline number says 90.7% of enrolled Australians voted at the 2025 federal election. That sounds like a system that works.
Flip it.
1,681,843 enrolled Australians did not vote. Another 333,396 were eligible but not even on the roll. And 919,512 who did show up spoiled their ballots so badly they did not count.
Add it up. 2.93 million Australians. Out of 18.4 million eligible. One in six whose voice did not reach the count. (Source: AEC 2025 Tally Room; AEC Enrolment Statistics 2025)
The state gap
Tasmania: 93.19%. Victoria: 92.44%. ACT: 92.41%.
Northern Territory: 70.83%.
Not close to the national average. Twenty points below it. The gap between the highest-turnout state and the lowest is 22 percentage points. Within a single country. (Source: AEC House Turnout by State, 2025)
| State | Turnout | Gap vs National |
|---|---|---|
| Tasmania | 93.19% | +2.49 |
| Victoria | 92.44% | +1.74 |
| ACT | 92.41% | +1.71 |
| NSW | 91.58% | +0.88 |
| South Australia | 91.47% | +0.77 |
| Queensland | 88.66% | -2.04 |
| Western Australia | 88.15% | -2.55 |
| Northern Territory | 70.83% | -19.87 |
Western Australia and Queensland sit below the national average too. Not by NT margins. But below. The gap is not just Territory-shaped. It is everywhere that is not a southern capital.
The electorate gap
The lowest-turnout electorates tell a sharper story.
Lingiari, which covers 1.35 million square kilometres of the NT, including 43% Aboriginal and Torres Strait Islander population: 62.21% turnout.
Durack, covering most of regional Western Australia: 78.31%.
Solomon, Darwin metro: 79.25%.
The highest: Deakin in Victoria at 94.89%.
The gap between Lingiari and Deakin is 32 percentage points.
Lingiari had 76,817 enrolled voters. 47,789 voted. 29,028 people did not vote in a single electorate. (Source: AEC House Turnout by Division, 2025)
The gap between Lingiari and Deakin is 32 percentage points. In one country. Under one compulsory voting system.
Why they do not vote
The best data comes from a Victorian Electoral Commission survey of 999 non-voters after the 2018 state election. No equivalent federal survey exists. The AEC sends 1.3 million failure-to-vote notices every federal election and has never asked a single one of those people why.
Victoria did. This is what they found.
87% said they were unavailable. 42% were overseas. 18% were at work. 16% were interstate. (Source: VEC Non-Voter Survey, 2018)
42% said they did not know. 28% did not know when the election was. 22% did not know it was happening at all. In a country with compulsory voting, where every enrolled voter gets a letter, more than two in five non-voters said they did not know the election was on.
30% said they were unwilling. 15% disliked the candidates. 14% objected to compulsory voting itself.
For remote Indigenous communities, the reasons stack differently. The AEC sent remote polling teams into the bush in 2022 with zero interpreters. Australia has more than 150 living Indigenous languages. Some remote communities get a single mobile polling visit lasting a few hours. If you miss that window, you do not vote.
NT turnout dropped 2.25 points from 2022 to 2025. Post-Voice referendum disillusionment is real. Indigenous Australians could not vote until 1962. Compulsory voting only applied to them from 1984. The JSCEM, the parliamentary committee that oversees elections, has repeatedly recommended the AEC publish estimated Indigenous turnout by division. The AEC has not done so. (Source: JSCEM Report on Indigenous Participation; ABC News)
The fine that time forgot
The penalty for not voting is $20.
It has been $20 since 1984. Forty-two years. Not indexed. Not adjusted. In 1984, $20 was a decent night out. In 2026, it is less than two coffees.
The maximum court fine is $330. At the 2025 election, the AEC sent 1,309,568 failure-to-vote notices. By the end of the 2022 cycle, only 226 people were convicted in court.
1.3 million notices. 226 convictions. An enforcement rate of 0.02%.
A Griffith University study found 47% of Gen Z who voted said they did so mainly to avoid the $20 fine. Not because they cared about the outcome. Not because they believed in the process. Because twenty dollars. (Source: Griffith University, March 2026; The Conversation)
47% of young voters citing the fine as their primary motivation suggests the compulsory model drives compliance more than engagement. Whether that matters depends on whether the goal is turnout or informed participation.
Would it change results?
Yes. The numbers say so.
The five closest seats in 2022:
| Seat | Margin | Non-voters | Ratio |
|---|---|---|---|
| Gilmore (NSW) | 373 | ~10,000 | 27x |
| Brisbane (QLD) | ~1,198 | ~12,000 | 10x |
| Deakin (VIC) | ~1,250 | ~10,000 | 8x |
| Chisholm (VIC) | ~2,300 | ~10,000 | 4x |
| Boothby (SA) | ~3,000 | ~10,000 | 3x |
Gilmore was decided by 373 votes. Ten thousand enrolled voters in that electorate did not vote. The non-voters outnumbered the margin by 27 to one.
The academic evidence is clear. Anthony Fowler, in a peer-reviewed study published in the Quarterly Journal of Political Science (2013), found that compulsory voting shifted Australian election results “drastically” toward the median citizen and toward left-leaning policy. Malkopoulou (2020) found that full turnout decreases radical party influence. (Source: Fowler 2013, QJPS; Malkopoulou 2020)
Non-voters are disproportionately young, lower-income, renters, and Indigenous. Census and survey data consistently show these groups favour Labor and the Greens at higher rates than the Coalition. The CIS “Generation Left” report found only one in four voters under 40 gave their primary vote to the Coalition. (Source: CIS, “Generation Left”)
If every non-voter voted, Labor and the Greens would likely gain seats. That is what the demographic data indicates. Whether that is a reason to act depends on whether the principle is maximum participation or the status quo.
The real number
The real participation rate is not 90.70%. That is turnout of enrolled voters who cast a valid ballot. It excludes everyone the system lost before the count started.
- 18,432,193 eligible Australians
- 333,396 not enrolled (1.8%)
- 1,681,843 enrolled but did not vote (9.3%)
- 919,512 voted but ballot was informal (5.0%)
Effective participation: 15,497,442 valid votes out of 18,432,193 eligible. That is 84.1%.
Not 90%. Not 91%. 84%. One in six eligible Australians did not cast a valid vote. (Source: AEC 2025 Results; Australia Institute)
What is not being measured
The AEC collects turnout data by electorate. It does not publish it by town, by postcode, or by socioeconomic status. You cannot look up whether your suburb votes at a higher rate than the one next door. The data exists at the booth level. The AEC does not make it easy to access in bulk.
Three specific data gaps matter.
No national Indigenous turnout rate. The AEC does not track Indigenous identity on the roll. Lingiari is the best proxy. The JSCEM recommended the AEC publish estimated Indigenous turnout. They have not.
No federal non-voter survey. Only Victoria has surveyed non-voters about reasons. The AEC sends 1.3 million notices a year and does not ask why. If the goal is to improve participation, understanding who is not participating and why would be a reasonable starting point.
No exemption data. The AEC does not publish how many people claim valid reasons for not voting, what those reasons are, or how many are accepted versus rejected. 1.3 million notices sent. No public reporting on the outcomes. (Source: AEC Non-Voters; AEC Annual Report)
The other side
84% effective participation is still high by global standards. The United States, with voluntary voting, manages roughly 66% in presidential years. Many European democracies sit in the 70-80% range. Australia’s floor is higher than most countries’ ceiling.
Compulsory voting itself is debated. Some democracies - New Zealand, the UK, Canada - argue that voluntary voting produces voters who actually want to be there. The counter-argument is that voluntary systems exclude the people who need representation most. Both positions have evidence behind them.
The $20 fine may be too low to deter non-voting, but increasing fines disproportionately affects low-income people. A single parent in Mount Druitt and a banker in Mosman face the same penalty. Raising it to $50 or $100 would hit the person already struggling to pay rent far harder. That is a policy trade-off, not a simple fix.
The AEC has limited resources. Surveying 1.3 million non-voters is not trivial. The Victorian survey covered 999 people. Scaling that nationally would require funding, methodology design, and years of baseline data. That does not mean it should not happen. It means it is not a quick win.
Some non-voters have legitimate reasons. They were overseas. They were in hospital. They were dealing with a family emergency. The exemption system exists for these cases and, by all available evidence, processes them routinely. The people who genuinely could not vote are not the problem.
The VEC data on awareness - 42% did not know the election was on - comes from a 2018 state election survey. It may not reflect federal reality. Federal elections receive more media coverage. The AEC runs national advertising campaigns. Whether the same unawareness gap exists at the federal level is unknown, because no federal survey has been done.
The way forward
Index the fine to inflation. Had the $20 fine been indexed since 1984, it would be roughly $55 today. That is still modest. It signals that non-voting has a consequence without making it punitive. Parliament could set this in a single sitting week.
Conduct a federal non-voter survey. Copy Victoria’s approach. Sample 2,000 non-voters after each federal election. Ask why. Publish the results. The AEC already has the contact details from failure-to-vote notices. The methodology exists. It needs funding and the political will to ask.
Publish estimated Indigenous turnout by division. The JSCEM has recommended this more than once. The AEC has not done it. Without the data, the gap stays invisible. Lingiari’s 62% turnout is a proxy, not a measurement. Every electorate with a significant Indigenous population deserves a number.
Expand digital enrolment and automatic enrolment. The AEC already cross-references tax and Medicare data. Using those datasets to automatically enrol eligible citizens - and to send reminders when elections are called - would address the 333,000 people who were eligible but not on the roll. Some states already do this.
Expand mobile polling for remote communities. More visits. Longer hours. Interpreter services in local languages. The AEC sent zero interpreters on remote polling teams in 2022. That is a fixable failure. Remote communities should not lose their vote because the polling team came once, for three hours, in a language they do not speak.
Sources
- AEC 2025 Federal Election Tally Room - Turnout by State
- AEC 2025 Turnout by Division
- AEC 2025 Informal Votes by State
- AEC Non-Voters
- AEC Enrolment Statistics 2025
- AEC Compulsory Voting Backgrounder
- VEC Non-Voter Survey 2018 (PDF)
- Fowler (2013) “Electoral and Policy Consequences of Voter Turnout” - QJPS
- Malkopoulou (2020) “Compulsory voting and right-wing populism”
- Australia Institute - 2022 Turnout Analysis
- CIS “Generation Left” Report
- Griffith University Gen Z Study
- The Conversation - Gen Z Fine Avoidance
- JSCEM Indigenous Participation Report
- Contreras et al. (2024) Electoral Studies
- Antony Green 2022 Pendulum
Related Investigations
Electoral and Democratic Failure
The $37 Million Election That Lost 600 Ballots
South Australia spent $37 million on its 2026 state election. Then 600 ballot papers turned up uncounted. A party volunteer was counting votes. The man in charge went on leave.
One Nation Banked $6 Million in Public Funding. Its Candidates Paid Their Own Way.
One Nation received more than $6 million from the AEC after the 2025 election for winning zero House of Representatives seats. Former candidates say they were left thousands of dollars out of pocket. The party refuses to say where the money went.
Who Owns the Ground Beneath You?
Eighty-six per cent of Australian mining is foreign-owned. BlackRock, Vanguard, and State Street control up to a third of each top miner. A Chinese state company holds the single biggest stake in Rio Tinto. Fortescue is the only one that is mostly Australian. Here is who actually owns the minerals under your feet.