Inferensys

Glossary

Quadratic Voting

A collective decision-making mechanism where participants express the intensity of their preferences by buying votes, with the cost of votes being the square of the number purchased, to balance majority rule against minority passion.
Cinematic overhead of a WeWork creative suite room with multiple curved monitors showing AI decision dashboards, executives in casual attire reviewing data, dramatic pendant lighting.
COLLECTIVE DECISION-MAKING MECHANISM

What is Quadratic Voting?

A mechanism for collective decision-making where participants express the intensity of their preferences by purchasing votes, with the cost being the square of the number of votes bought, balancing majority rule against minority passion.

Quadratic Voting (QV) is a collective decision-making mechanism where participants allocate 'voice credits' to purchase votes on issues, with the cost in credits being the square of the number of votes cast. This quadratic cost function forces voters to make a concrete, measurable sacrifice to express the intensity of their preference, preventing a tyrannical majority from completely ignoring a deeply invested minority.

The mechanism addresses the fundamental limitation of traditional one-person-one-vote systems, which capture preference direction but not intensity. By making the marginal cost of each additional vote increase linearly, QV achieves an optimal resource allocation equilibrium where the collective outcome reflects the greatest aggregate utility, a principle derived from marginal utility theory in economics and formalized by Glen Weyl and Steven Lalley.

MECHANISM DESIGN

Key Features of Quadratic Voting

Quadratic Voting (QV) is a collective decision-making mechanism designed to balance the tyranny of the majority against the intensity of minority preferences. It achieves this by making the cost of votes increase quadratically rather than linearly.

01

Quadratic Cost Function

The defining mathematical rule of QV: the total cost to cast n votes on a single issue is the square of the number of votes purchased (). This means 1 vote costs 1 credit, 2 votes cost 4 credits, 3 votes cost 9 credits, and so on. This non-linear pricing forces participants to economize, allocating more votes only to issues they feel most passionately about, thereby revealing the intensity of preference, not just the direction.

02

Sybil Resistance via Identity

QV's mathematical properties depend on one-person-one-credit allocation. Without robust identity verification, an attacker could create multiple pseudonymous identities (a Sybil attack) to buy votes cheaply at the linear start of the cost curve, circumventing the quadratic penalty. Effective implementations require a strong Sybil resistance mechanism, often using Soulbound Tokens, Verifiable Credentials, or biometric proof-of-personhood protocols to ensure each human receives an equal, non-transferable pool of voice credits.

03

Optimal Preference Revelation

Under standard economic assumptions, QV is a theoretically optimal mechanism for public goods funding. A rational voter will allocate credits such that their marginal vote cost equals the marginal benefit they derive from the decision. This results in the collective outcome approximating the one that maximizes aggregate utility. In practice, this is used in Quadratic Funding for open-source software, where a matching pool amplifies the number of individual contributors, not just the total sum donated.

04

Collusion Resistance and Limits

A primary vulnerability of QV is collusion, where a coordinated group of voters pools their credits to act as a single, outsized influence. This breaks the assumption of independent decision-making. Mitigation strategies include pairwise coordination subsidies or algorithmic detection of coordinated voting patterns. The mechanism is also susceptible to the tyranny of the majority if a large, apathetic group votes uniformly against a small, intense minority on a low-stakes issue for them.

05

Vote Delegation and Liquid Democracy

QV can be combined with Liquid Democracy to reduce voter fatigue. Participants can delegate their voice credits to a trusted expert or representative on specific topics while retaining the right to override that delegation and vote directly. This creates a dynamic, issue-based representation system. The delegated credits are still subject to the quadratic cost rule when the delegate casts them, preserving the intensity-weighting property of the mechanism.

06

Real-World Implementations

QV has moved beyond theory into practical governance. The Colorado Democratic Party used it in 2019 to prioritize platform planks. Gitcoin uses Quadratic Funding to allocate millions in matching funds to Ethereum ecosystem public goods. The city of Nashville, TN, has experimented with QV for participatory budgeting. These trials demonstrate its viability for aggregating community sentiment, though user education on the non-linear cost curve remains a critical adoption challenge.

QUADRATIC VOTING

Frequently Asked Questions

Explore the mechanics, game theory, and practical applications of Quadratic Voting, a collective decision-making mechanism designed to balance majority rule with the intensity of minority preferences.

Quadratic Voting (QV) is a collective decision-making mechanism where participants allocate votes to express the intensity of their preferences, rather than just the direction. The core mechanism dictates that the cost to a voter is the square of the number of votes purchased. For example, casting 1 vote costs 1 token, but casting 5 votes costs 25 tokens. This quadratic cost function forces participants to economize their influence, ensuring that a highly passionate minority can outbid a disinterested majority only if the issue is truly critical to them. The system mathematically aggregates these weighted preferences to find the option that maximizes overall welfare, preventing the tyranny of the majority while still respecting broad consensus. It was formalized by economist Glen Weyl and legal scholar Eric Posner in their book Radical Markets.

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.