Decentralized Governance Models for Blockchain-Based Organizations

Abstract

Decentralised Autonomous Organisations (DAOs) manage over USD 28 billion in treasury assets as of mid-2025, yet most operate under governance models designed ad hoc by founding teams with limited political science or mechanism design expertise. The consequences are visible: voter apathy (median turnout below 4% on Snapshot), plutocratic capture (whales controlling majority vote weight), governance attacks (flash loan-funded vote manipulation), proposal fatigue (hundreds of proposals with declining engagement), and delegate concentration (fewer than 20 delegates controlling majority power in protocols like Uniswap, Compound, and Arbitrum). These failures are not inevitable -- they reflect design choices that can be evaluated, compared, and improved. We present the Decentralised Governance Assessment Framework (DGAF), evaluating six governance models -- token-weighted voting, quadratic voting, conviction voting, futarchy (prediction market governance), reputation-based governance, and optimistic governance -- across five governance dimensions: participation rate, plutocracy resistance, attack resistance, decision quality, and operational efficiency. Our Governance Model Effectiveness Score (GMES) evaluates each model on 14 real DAO deployments with over 12,000 governance proposals and 840,000 individual votes. Quadratic voting achieves the highest GMES (0.904) by providing the strongest plutocracy resistance (0.940) while maintaining practical participation rates, while optimistic governance achieves the highest operational efficiency (0.960) by requiring active engagement only for contested decisions.