Liquid Democracy: Governance After Red vs. Blue

The case for Liquid Democracy, and the design choices that determine whether it WILL actually work.

I’ve long been interested in upgraded governance systems. Some of these ideas first appeared in three essays I published on my Medium in 2018 and 2019. The conditions that motivated them have only worsened. This is the consolidated, updated, deeper and considerably more sober version.

American democracy runs on 18th-century governance hardware. The Founders designed it for a population OF 3 MILLION, that traveled at horse speed, communicated by pen and ink, and assembled in rooms heated with fireplaces. The Constitutional system was extraordinary for its moment. It is now eight technological revolutions out of date.

Two teams own the field. The match between Team Red and Team Blue is fixed, not in the sense that either side throws it, but in the sense that the rules guarantee a stalemate that benefits both teams’ professional class. Congress sits near 10 percent approval, lower than used car salesmen. Younger Americans have begun to question democracy itself. Infrastructure rots. Debt compounds. Climate change runs its experiment on us in real time with no serious response.

This essay sketches an alternative. Not a revolution, not a constitutional rewrite, but an upgrade compatible with the existing system: Liquid Democracy. What follows: what it is, why it works, the honest difficulties, the design questions most discussions duck, the constitutional path to introducing it, a hard look at prior attempts, and the funding mechanism that determines whether the proposal survives contact with reality.

The shared-values bedrock

Start with what most Americans actually share. Strip the partisan packaging from the values themselves and the overlap is enormous:

• Safety: national defense, secure neighborhoods, clean water, air, food.

• Equal opportunity: no arbitrary barriers based on race, religion, gender, origin, orientation.

• Fairness: hard work and good character get rewarded; cheating and free-riding do not.

• Social insurance: families that fall on hard times have somewhere to turn.
  
  

• Future generations: their education, health, and character.

• Sustainability: we leave a world worth inheriting.

• Freedom: people speak their minds, make their own choices, take their own risks.

Pluralism guarantees disagreement about priority and method. Spend on defense or healthcare? Public schools or charters? These are real questions, worth real argument. But the bedrock values are widely shared. A working political system would source from those values and operate in good faith to give most people most of what they want most of the time.

Our current system does not do that. It cannot, given how it is structured.

Why the current system fails

The roots run deeper than the personalities involved. Winner-take-all elections combined with low-turnout primaries dominated by the most polarized voters guarantee that the most extreme partisans drive the nominating process. Add the revolving door, in which congressional staffers cash out as lobbyists and retired officials join boards, and you get a professional political class whose career incentives point away from solving problems. Solving problems does not pay. Posturing does.

The result is predictable: gridlock on substance, theater on hot-button issues, a steady drip of patronage to the connected. Congress has not passed all the funding bills required to run the government on time since fiscal year 1997. National debt now grows during expansions, not just recessions. The two-party duopoly is not a bug in this system. It is the system.

Liquid Democracy: the mechanism

Liquid Democracy is a technologically enabled hybrid of direct and representative democracy. The core idea is simple: every voter retains the right to vote directly on every issue. Most voters will not have the time or interest. So every voter also has the option to delegate the vote, give a proxy, to a person or organization they trust.

Four properties make it work:

1. Every voter can vote directly on any issue, any time.

2. Every voter can delegate the proxy to a chosen delegate: an individual, a non-profit, an advocacy group.

3. A delegate who receives a proxy can pass it further to another delegate. Proxies chain.

4. At any time, a voter can revoke the proxy and either reassign it or vote directly.
   
   

To keep cognitive load manageable, all legislation gets divided into roughly 25 issue areas: defense, environment, civil liberties, healthcare, taxes, energy, education, and so on. A voter might give her defense proxy to her cousin the Air Force major, her environment proxy to the Sierra Club, her healthcare proxy to a physicians’ organization, and vote directly on civil liberties because that is where her own conviction is strongest. If the cousin starts voting in ways she dislikes, she pulls the proxy and reassigns it. The proxy is liquid by design.

The Sierra Club, having received millions of environment proxies, might further delegate them, passing some to a wildlife-focused affiliate and retaining others for direct vote. The Club might also consult an advisory committee of climate scientists. The structure is recursive.

Four benefits worth the effort

Better decisions. Specialized delegates carry specialized knowledge. A debate about procuring 120 fighter aircraft would be shaped by Air Force veterans, defense-procurement watchdogs, and budget hawks. These are interest groups operating in good faith, not lobbyists representing the prime contractor. The decision-making process gets its expertise back.

More engagement. More than a third of voting-age citizens skipped even the 2024 presidential election. Of those who vote, most cannot name their congressional representative. The current system signals to most voters that their participation is symbolic at best. Liquid Democracy reverses the signal: every voter has a precise, granular, reassignable stake.

Better alignment. The socially-progressive-fiscally-conservative voter is currently homeless. She must pick a team that misrepresents her on half the issues she cares about. Under Liquid Democracy she splits her proxies: progressive delegates on social issues, fiscally conservative delegates on economic ones. The mismatch shrinks dramatically. Voters stuck in dominant-party states find their preferences winning some of the time instead of none.

A reshaped geometry for money in politics. Today there are 535 members of Congress to lobby. Under Liquid Democracy there are millions of delegates across many issue areas. The math of influence changes from buying access to a few hundred people to attempting to shape millions. Money does not disappear from politics. It migrates into delegate-discovery operations, platform-influence campaigns, proxy-broker networks, celebrity-delegate cultivation, and recommendation-algorithm shaping. But the geometry of corruption shifts from concentrated to dispersed, and dispersed corruption is harder to maintain, easier to expose, and structurally weaker.

Three honest challenges, each solvable in principle

Mass-psychology runaway. Human beings are perfectly capable of mass delusion. Imagine major civil-liberties legislation voted on the week after September 11, 2001. The cure is viscosity: the platform should slow things down deliberately. Mandatory waiting periods between bill introduction and vote. For major legislation, a required second confirming vote 30 or 60 days after the first. Time for voters to reconsider, reassign, reclaim. Time for the mob mood to fade.

Tyranny of the majority. The Founders had reason for a Bill of Rights, and Liquid Democracy must preserve equivalent protections. Some classes of legislation, those restricting civil liberties, declarations of war, constitutional amendments, should require supermajorities. Viscosity plus supermajority requirements approximate the deliberative slowness that the Constitution itself was designed to impose.

The software question deserves a separate honest treatment, and it is no longer the least daunting part. A national Liquid Democracy platform would face phishing, identity theft, coercion, denial-of-service attacks, insider manipulation, foreign influence operations, deepfake delegate endorsements, bot-driven persuasion campaigns, and attacks on the proxy graph itself. Open-source platforms have been built and tested in Europe, but none has had to survive nation-state adversaries operating at scale. The software problem is tractable only if security, privacy, and auditability are treated as constitutional-level design requirements from the start. Tractable does not mean easy.

Who controls the platform

The deepest design problem is not the mechanism. It is the meta-mechanism: who governs the system that governs the votes.

Whoever defines issue categories, drafts ballot wording, sets vote timing, manages identity verification, runs delegate-search and recommendation algorithms, handles dispute resolution, and controls audit access holds enormous structural power. A Liquid Democracy platform with the wrong governance is a more efficient capture mechanism than the system it replaces. “Open source and auditability” is necessary but not sufficient. The platform’s own governance is the first political question, not an afterthought.

The privacy-versus-accountability tradeoff is its own design fork, with no clean answer. If proxy assignments and votes are public, accountability is strong but voters become targets for employer pressure, family coercion, harassment, and ideological blacklisting. If they are private, voters are protected but auditing, delegate accountability, and proxy compensation all become harder. The German Pirate Party went transparent and got a years-long privacy backlash that nearly broke the experiment. Cryptographic schemes that preserve vote secrecy while permitting verification, and pseudonymous proxy chains with selective disclosure for compensation auditing, are real research areas with real progress. They do not solve themselves.

Whatever the architecture, platform governance must be designed for adversarial conditions, structured so that no single actor, including the state, can quietly change the rules. This is constitutional-grade work. It deserves constitutional-grade scrutiny. Anyone proposing a Liquid Democracy platform without a serious answer to “who governs the governor” is selling something.

The constitutional path

The U.S. Constitution does not authorize a digital direct-democracy override. Congress votes through its elected members. But nothing in the Constitution prevents an elected member from pledging, as a condition of seeking office, to vote according to her constituents’ instructions as registered on a Liquid Democracy platform.

That pledge is the wedge. A candidate runs for any office, city council to U.S. Senate, and commits to vote per the platform’s output for her constituency. She remains a constitutional legislator. The platform becomes her conscience and her instruction set.

The pledge is politically binding, not legally enforceable. Holding pledge candidates to their commitments depends on platform transparency, constituent vigilance, and electoral consequence. This is workable, not bulletproof.

A pledge candidate must handle the rest of legislative work: committee service, drafting, negotiation. She polls her Liquid Democracy constituency on negotiating positions, returns compromise language for ratification, and refuses to vote on rushed legislation that gives the platform no time to deliberate. The 1,400-page bill dropped at midnight and voted on at dawn defaults to No. Liquid Democracy office holders can stand up advisory committees of acknowledged experts, who themselves can act as delegates within their domains.

The structure scales. Start with city councils and state legislatures. Generate visibility, build platform maturity, prove the model. The federal stage follows.

The 51/49 problem

A pledged representative still casts a single indivisible vote on the floor. The platform speaks in proxy weights and continuous tallies; the constitutional system speaks in Yes or No. What happens when the district’s platform splits 51 to 49?

The immediate rule has to be strict simple-majority pass-through. The platform says 51 percent Yes, the representative votes Yes. The representative is a transmission belt for the aggregate will of her constituency, not a tiebreaker exercising independent judgment. Anything else turns the pledge into a fig leaf.

This is lossy compression. A 51/49 split is not the same political signal as a 95/5 split, and rounding both to “Yes” throws away information the platform worked to gather. It is the necessary compromise required to run next-generation governance software on 18th-century hardware. Fit the current constitutional arrangement, then transcend it.

The long-term trajectory points toward fractional voting. Once the pledge mechanism proves itself at scale, the next objective is modifying legislative rules, and eventually constitutional frameworks, to let representatives cast fractional votes on the floor. In a mature system, a 51/49 district registers exactly as 0.51 Yes and 0.49 No, tabulated alongside the fractional outputs of every other district. The lossy compression goes away. The exact mandate of the electorate becomes the actual measurement on the floor.

That is not a 2026 reform. It is the natural endpoint of a working pledge-candidate system. The pledge gets us in the door. Fractional voting is what the room looks like once we own it.

A short history, and an honest reckoning

The idea is older than the software. Lewis Carroll proposed proxy-like voting mechanisms in 1884. The Oregon reformer William S. U’Ren pushed something he called interactive representation in 1912. Gordon Tullock floated transferable-vote ideas in 1967, James C. Miller in 1969. The first internet-era articulation came from Christopher Lanphier in 1995. An online theorist using the handle Sayke coined the term “Liquid Democracy” around 2000. The canonical modern formulation is Bryan Ford’s 2002 paper “Delegative Democracy,” which laid out transitive proxy with revocability and remains the seminal reference.

The concept is well over a century old in its proxy-voting form, and a quarter-century old in its current digital form. It has been attempted multiple times in real political settings. None of the attempts has broken through. Worth a hard look at why.

Demoex, Sweden (2002 to 2014). A philosophy teacher and a handful of high school students in Vallentuna, a Stockholm suburb of about 30,000, formed a party committed to direct-democracy voting through an online platform. Their elected representative was bound to vote in council according to the platform’s outcome. In twelve years on the local council, Demoex developed roughly 80 proposals, brought 33 to the council, and got 6 approved with modifications. The party folded around 2014 when its bid for a national-level role failed and its underlying software vendor went bankrupt. A genuine accomplishment at small scale. No upward trajectory.

The German Pirate Party and LiquidFeedback (2009 onward). The most ambitious real-world test. LiquidFeedback, the open-source platform that came out of the Pirate movement in 2009, implements transitive delegation cleanly. The Pirates entered four German state parliaments in 2011 and 2012. They peaked at roughly 35,000 members and 13 percent in national polls. Then collapse. Berlin Pirates went from 8.9 percent in 2011 to 1.7 percent in 2016, losing all 15 seats. The 2025 federal election: 13,800 second votes, about 0.03 percent. Terminal. The software survives, picked up here and there by German counties, a few corporations, and Slow Food Germany. The party that built it is gone in any meaningful sense.

What killed it was not Liquid Democracy. What killed it was politics: amateur candidates who could not handle the substance, internecine warfare that consumed the movement’s energy, and radical transparency taken so far that internal bickering ended up as press copy. The tool worked. The political organism using it did not.

The American 2017 to 2019 pledge wave. Camilo Casas ran in Boulder. David Ernst and Sergey Piterman in California. Brian Lehman in Colorado. Matthew Wagoner in Missouri. Kurtis Hanna had run earlier as a Republican using LiquidFeedback for Minnesota’s House District 62a in 2012. None broke through. None generated a successor wave. The platforms were primitive. The candidate pool was thin. The mainstream press treated the idea as a curiosity rather than a serious proposal.

DAO governance, roughly 2020 to now. Token-weighted “liquid delegation” is now a dominant pattern in the crypto governance space. The results are sobering. One-token-one-vote produces plutocracy reliably. Delegation patterns concentrate. Voter fatigue is endemic. Whales accumulate delegations. Vitalik Buterin himself has publicly warned that simple coin voting tends to break in predictable ways. Academic reviews of the DAO experience describe the same failure modes political theorists predicted: collusion, vote-selling, abstention exploitation, delegate capture. The crypto space is the largest live experiment in transitive delegation that has ever run. The verdict so far is that the mechanism alone is not enough.

What the record shows.

The mechanism works at small scale and in tightly-bounded communities. It does not yet scale to legitimate democratic governance.

The software is the easy part. Legitimacy, identity verification, the privacy-versus-accountability tradeoff, and political maturity of the participants are the hard parts.

Funded delegates are not optional. Every real-world attempt has expected delegates to work for free or for ideological satisfaction. That fails in two predictable ways: burnout, or capture by paying interests. The funding mechanism described next has never been tried at any meaningful scale. Until it is, the question of whether Liquid Democracy can work with adequately resourced delegates remains open.

Sympathy alone does not produce momentum. Demoex generated press. The Pirates generated press. The 2018 American candidates generated press. The DAOs generate continuous press. None of it has converted to durable political power. The audience exists. The vehicle has not been built.

This will not be easy. The barriers are not theoretical. They are technical, political, financial, and cultural. None is fatal. All are real. Anyone advocating for Liquid Democracy who pretends otherwise should be ignored.

The design choice everyone misses: pay the delegates

Here is the hard problem most Liquid Democracy advocates skip. Being a competent delegate is real work. An education-policy delegate who wants to deserve a proxy must follow research, read legislative drafts, attend hearings, build relationships, sometimes hire staff. Multiply that across 25 issue areas and you have a serious labor input across the system.

If that work goes uncompensated, two failure modes follow. Either delegates are unpaid hobbyists who cannot sustain the work and burn out, or delegates get paid by interests with axes to grind. The second mode is just lobbying with extra steps. Either way the system collapses back into the failure it was meant to escape. This is exactly what every prior real-world attempt has demonstrated.

The fix: a Democracy Allowance.

The proposal: every registered voter receives a monthly allowance of $5 in Delegate Dollars from the U.S. Treasury. $60 per year. With more than 211 million active registered voters, the program costs roughly $12.7 billion annually. For scale: the 2024 federal election cycle was projected at $15.9 billion. The legislative branch costs about $7 billion to operate each year. The allowance is in the same order of magnitude. Adjust up or down as experience dictates.

The mechanism: each voter’s allowance flows proportionally to whoever holds her proxies. 25 issue areas, so each area carries 20 cents per month. A voter who proxies all 25 areas to one delegate sends that delegate $5 monthly. A voter who proxies one area to Delegate A, ten to Delegate B, and retains the rest sends 20 cents and $2 respectively; the unallocated remainder returns to the Treasury. If a delegate further chains a proxy, the money chains with it.

The economics work out interestingly:

• A professional delegate with 1,000 full proxies earns $60,000 a year. A living wage in much of the country.

• A part-time delegate with 150 full proxies earns $9,000 a year. Enough to incentivize a political hobbyist who wants to serve neighbors, family, and community better than they would serve themselves.

• The Sierra Club holding a million environment proxies generates a $2.4 million annual budget for that issue area alone. Enough to staff serious research and legislative drafting. Enough to compete with industry.

• A million voters proxying fully to a minor party (Greens, Libertarians, or whatever new formations emerge) generates $60 million per year. Several times the current budget of any third party. Alternative political visions get oxygen.

A few design questions worth flagging:

Who is eligible to receive proxies? A workable answer: anyone can receive proxies up to a low monthly cap, perhaps $5,000. Beyond that, formal registration with background disclosure and area specification is required. As well as financial disclosures.

Linear or non-linear formula? A non-linear formula could penalize delegates who hold themselves out for all 25 areas, encouraging specialization. Worth testing.

Celebrity delegates: a problem? Yes. Any power-law attention system will produce celebrity delegates. Some will professionalize: hire policy staff, publish reasoning, run issue-specific advisory boards, and become useful public-facing aggregators. Others will become fronts for organized interests, proxy farms, brand vehicles, or vanity projects. The system should not pretend celebrity delegation is harmless, nor should it try to ban it. The design response is disclosure, auditability, proxy-concentration monitoring, conflict-of-interest reporting, fiduciary-style obligations and reporting for high-volume delegates, anti-kickback rules, and effortless revocation. Voters should know who funds a delegate, who advises the delegate, how the delegate has voted, whether proxy growth looks organic or engineered, and whether the delegate is passing proxies onward. Celebrity delegation is a predictable attack surface. Treat it that way.

Fraud, capture, and abuse. A $12.7 billion annual program attracts predators. The system needs hardening against fake delegates, shell organizations, proxy kickbacks, foreign-funded influence operations, employer and union coercion of rank-and-file voters, and delegate-farming operations. A workable framework: mandatory disclosure of beneficial ownership for organizational delegates, anti-kickback rules with criminal penalties, foreign-funding bans, conflict-of-interest reporting, proxy-concentration alerts that flag suspicious accumulation patterns, and independent audit authority with subpoena power. The Democracy Allowance is a public program. It needs public-program-grade integrity controls. None of this is exotic. All of it has analogs in existing campaign finance, securities, and tax-exempt-organization regulation. The lift is legal-drafting work, not invention.

One more point. Think of the Democracy Allowance not as a budget line but as a Prigogine-style energy flow: new and complex structures emerge wherever energy moves through a system far from equilibrium. $12.7 billion per year flowing through a recursive proxy graph, with constant reassignment and constant re-evaluation, will not produce the static equilibrium of today’s lobbying class. It will produce evolution. Novel forms of political representation we cannot now envision. The flow opens the search space for how to do politics.

What to do now

Liquid Democracy is not a finished product. The platforms need work. The legal frameworks need building. The candidate pool needs cultivation. The pledge mechanism needs road-testing in real elections. Auditability, transparency, and security must be engineered as first-class requirements, not afterthoughts. Platform governance is its own constitutional problem and deserves its own design effort.

None of this is fantasy. We have built harder things. The barrier is not technical or constitutional. The barrier is learned helplessness, the conditioned belief that nothing we do could change the underlying game.

A few concrete moves:

• If the argument lands, spread it. Email, talk, share. The first job is making the concept legible to enough people that pledge candidates have a constituency to run for.

• Build or join a platform. Open-source Liquid Democracy software exists. It needs hardening, scaling, security engineering, and platform-governance design. Researchers, engineers, and policy people are welcome.

• Run. City council, school board, state assembly. The pledge mechanism works at any scale and the low-stakes races are where the model gets debugged.

• Fund the work. Donate to pledge candidates. Donate to platform builders. Make the case for the Democracy Allowance.

Demoex was small. The Pirates collapsed. The 2018 American wave did not break through. The DAOs are the wrong shape. Each attempt failed in its own specific way, and each failure narrows the design space for the next attempt. Better-funded delegates. Hardened platform governance. Security-first engineering. Abuse-resistant funding. Pass-through pledges now, fractional voting later. These are not optional features. They are the conditions for the mechanism to function at all.

Two teams have run the field for a long time. They are visibly failing. The audience knows.

Liquid Democracy is a sketch of a different sport. Improve the sketch. Build the platforms. Run the candidates. Pay the delegates. Govern the platform that governs the votes.

Reclaim the game.