What if every voter had $50 to give a candidate? How to fix money in Australian politics
Donation caps, fast disclosure, public funding, open diaries, and a voucher every voter spends. Seattle has run one since 2017. Here's what happened.


Taiwan lets anyone help write the rules for new technology. People post short statements online, everyone votes agree or disagree, and software finds the statements that people on opposite sides both accept. Those become the starting point for the law. When Taiwan used it to settle a fight over Uber, the government wrote every point of agreement into the new regulations.
Taiwan lived under martial law from May 1949 to July 1987, more than 38 years (Wikipedia, 2026d). Its people first chose their president directly in 1996 (Wikipedia, 2026a). Many Taiwanese adults remember a time when the government decided and the people obeyed.
That memory shapes what came next. In March 2014 the governing party tried to push a trade agreement with China through parliament without the clause-by-clause review it had agreed to. Students occupied the Legislative Yuan, Taiwan's parliament, from 18 March to 10 April: 23 days. On 30 March, police counted 116,000 people at the rally outside; organisers said 500,000. The Speaker promised no review until a law existed to monitor every cross-strait agreement, and the students left. The agreement was never ratified (Wikipedia, 2026f). This became known as the Sunflower Movement.
Some of the help came from g0v (pronounced "gov zero"), a community of civic hackers that started in late 2012. The zero stands for rethinking government "from zero" (Wikipedia, 2026b).
Their idea borrows from how programmers share code. Take a government website, swap the "o" in its address for a zero, and you reach g0v's "forked" copy: the same information, rebuilt so people can actually use it. If the copy is better, the government can "merge" it back into its own site (Wikipedia, 2026b). One of the first projects began when the government released a vague video about its economic plan. Instead of complaining, volunteers built Budget Maps, which show where tax money goes.
That's the cultural insight. g0v doesn't wait for permission, and it doesn't just protest. It builds the better version and offers it back.
Out of that community came vTaiwan, launched in 2014: an open process where citizens, experts and officials work through a national issue together, online and in person (Hsiao et al., 2018). It runs in four stages (vTaiwan, n.d.):
Most online discussion rewards the loudest, angriest post. Pol.is is designed the other way. Nobody can reply to anyone. Participants stay anonymous. You can post a statement, and you can vote agree, disagree or pass on other people's. That's it (Wikipedia, 2026e).
The software then groups people by how they vote and draws a map of the groups. It shows where the divides are, and, more usefully, which statements win support from every group at once (Horton, 2018). Because those are the statements that rise to the top, people start writing statements the other side can accept.
Uber's arrival in Taiwan split opinion into two camps, for it and against it. vTaiwan ran it through the process. More than 4,000 people took part in the Pol.is survey over four weeks (vTaiwan, n.d.).
At the start the voters fell into those two camps. As the statements piled up, some drew support from both, including (Horton, 2018):
The argument moved from "for or against Uber" to safety and fairness. Uber accepted the recommendations, and the government wrote all the Pol.is consensus points into its new regulations (vTaiwan, n.d.).
Audrey Tang, a core member of g0v, volunteered during the Sunflower occupation to broadcast the protesters' message. In 2016 Taiwan appointed Tang a minister without portfolio for digital affairs, and in 2022 the first Minister of Digital Affairs, a post Tang held until May 2024. Tang was the first transgender and non-binary official in Taiwan's cabinet (Wikipedia, 2026c).
Every one of Tang's ministerial meetings was recorded, transcribed and published online. During COVID-19, Tang's office met rumours with "humour over rumour": an answer within 20 minutes, in 200 words or fewer, with two funny pictures (Wikipedia, 2026c). The same idea runs through all of it. If you trust people with the facts, they tend to earn it.
vTaiwan reports that more than 28 cases have gone through the process and 80% led to decisive government action (vTaiwan, n.d.). That's a strong record. It also has real limits (Horton, 2018):
Taiwan's answer to the reach problem was a second platform. Join, the government's own national site for online deliberation, also uses Pol.is, and by 2018 nearly five million people had signed up (Horton, 2018; Wikipedia, 2026e).
None of this makes vTaiwan a failure. It's an experiment that works best on new, narrow questions, where nobody has dug in yet.
Not at national level. Australia's closest practice is offline: councils hand a hard question to a panel of residents picked by lottery. In Manningham, a Victorian council of 123,000 people, 40 residents chosen that way met seven times over three to four months to answer "What range and level of services are you prepared to pay for?" on a $147 million budget. They wrote their own report, and the council committed to implement it "to the maximum extent possible" (newDemocracy Foundation, 2024).
That reaches the same place from a different direction. A lottery panel is small but looks like the town. Pol.is is open to everyone but only reaches the people who show up. Taiwan shows a government can use both. Switzerland goes further still and lets citizens vote on the law itself; see How Swiss direct democracy works.
Donation caps, fast disclosure, public funding, open diaries, and a voucher every voter spends. Seattle has run one since 2017. Here's what happened.

Five per cent of donors gave over half the big parties' declared donations. What the records show about access in Canberra, and what has changed since.

Quadratic voting lets you show how much you care, not just which side you're on. How it works, where it's been tried, and what can go wrong.
