Liquid Democracy
Published May 12, 2025
Combining elements of direct and representative voting, liquid democracy allows individuals to either vote directly or delegate their vote to someone they trust. Real-world applications in blockchain governance, shareholder decision-making, and AI oversight reveal both structural challenges — low participation, vote concentration, and weak civic incentives — and signs of institutional promise. As digital platforms reshape how decisions are made, liquid democracy invites a broader reconsideration of how democratic systems might function in a more networked, decentralized society.
Check out more from Andrew B. Hall:
- Read "Harmonizing Democracy and Tech" by Andrew B. Hall here.
- Read "What Happens When Anyone Can Be Your Representative?" research paper by Andrew B. Hall here.
- Read "What Kinds of Incentives Encourage Participation in Democracy?" by Andrew B. Hall here.
Learn more about Andrew B. Hall here.
__________
The opinions expressed are those of the authors and do not necessarily reflect the opinions of the Hoover Institution or Stanford University.
© 2025 by the Board of Trustees of Leland Stanford Junior University.
>> Andrew B. Hall: Thomas Jefferson famously noted that a well-informed citizenry is not only the best defense against tyranny, but is the heart of a dynamic democracy. American democracy today operates much like it did at our nation's founding. Citizens periodically vote directly for representatives who then make decisions at the state, local, or federal level on their behalf.
But what if, as a voter, you don't consider yourself well enough informed? Today, your only choice is to vote anyway, to become more informed before you vote, or to abstain from voting altogether. But there's another option that might be worth exploring, liquid democracy. Liquid democracy is a concept dating back to the 19th century, but it has gained traction in recent years as our lives have become more digital.
In liquid democracy, voters retain the option to vote for themselves, but those who feel insufficiently informed can give their vote to someone else to cast on their behalf. It's a hybrid between representative democracy and direct democracy. And digital technology makes it feasible to build at scale for the first time.
Today, a variety of organizations are actively experimenting with different ways to delegate decision-making power to others. Major investment firms like BlackRock, State Street, and Vanguard now allow shareholders to delegate their votes through voting choice programs in blockchain. Token-based systems allow people to vote digitally to buy, sell, or lend voting power and to delegate their votes to others.
AI companies, including Anthropic, Meta, and OpenAI, are running public experiments to delegate decision-making power to randomly sampled citizen assemblies to provide guidance on what AI chatbots should and shouldn't say or what they can and cannot do. The real-world applications for liquid democracy could have potential. In California, for example, ballot initiatives, which are chances for voters to make decisions together outside the legislative process, can be so numerous that it's difficult to be informed on each one.
Many voters decline to vote on many of them or vote somewhat randomly. Liquid democracy could allow voters to give their voting power to someone more informed, which might help to push through the best initiatives and vote down the bad ones. While the idea of delegating votes might sound attractive in the real world, especially if it might lead to a better informed and more civic-minded citizenry, theoretical research suggests there could be significant challenges too.
To figure out whether liquid democracy works well and how it could work better, we need to test it out in a real-world, high-stakes setting and see what actually happens. Fortunately, blockchain startups are actively doing this, and we can learn from what they've experienced. One of the biggest concerns is preventing too much voting power from concentrating in a small number of people, so-called superdelegates.
Even if voters can reclaim their individual vote back. It risks concentrating too much unaccountable power in the hands of the few. Finding the right delegate can be its own challenge as well. Making informed decisions about who should be the next US President, US Senator, state senator, governor, sheriff, county clerk, or town mayor is without a doubt a lot to ask.
But finding someone informed to whom you can delegate your vote may be equally as hard, which might explain one of liquid democracy's other significant challenges. Our study is the first research project in political science to leverage these natural experiments in liquid democracy that are going on today in blockchain, collecting comprehensive data on liquid democracy and voting that covers over 250,000 voters and more than 1,700 proposals across 18 of the largest Ethereum-based DAOS.
We find both challenge and promise to liquid democracy first. In blockchain-based organizations that use liquid democracy, only 17% of voting power gets delegated, similar to turnout in many local elections. And those with delegated power often fail to vote, with delegate participation rates around 33%. This points to a potential problem.
Even when given more flexible voting options, many people still choose not to participate. Low overall rates of participation contribute to a second challenge. Voting power ends up concentrated in a relatively small number of superdelegates, threatening to undo the decentralization of these projects if voters do not monitor their delegates closely.
But there are also signs of promise. We find that delegation is bottom up more than top-down. Smaller token holders are more likely to delegate their votes than larger token holders, which is consistent with the hope that liquid democracy can induce people who have weak incentives to pay attention and participate to still have their votes counted.
Delegates with a better track record of participating attract more votes, suggesting that token holders are not delegating their votes entirely at random. Moreover, there's evidence that thoughtful design can help address these challenges. Delegation programs that create convenient online user interfaces and offer information on potential delegates, including their voting positions and their track records, show substantial increases in participation, increasing overall rates of delegation by roughly 25%.
The world is undergoing tremendous technological change, and it would be strange to think we cannot harness these changes to improve the way we govern ourselves. But the challenge is figuring out which ideas are genuine improvements and which are ill-advised. Although the future of liquid democracy and its applicability into real-world civic engagement remains uncertain for the time being, what is clear is that liquid democracy and the idea of delegated voting may become increasingly significant in the digital age.
Whether it's financial institutions, online platforms, or perhaps online governance, liquid democracy offers both promising possibilities and cautionary lessons about the challenges of translating democratic ideals into practical reality.