DAO Governance Models Compared: Token Voting, Optimism's Bicameral Model, and Security Councils

DAO Governance Models Compared: Token Voting, Optimism’s Bicameral Model, and Security Councils

Governance architecture is one of the most consequential and least well-understood dimensions of protocol design. Token distribution, smart contract security, and liquidity mechanisms receive enormous engineering attention. Governance mechanisms — the processes by which token holders collectively make decisions — often receive inadequate design attention and are retrofitted from templates without sufficient adaptation to the specific protocol’s circumstances.

The results are visible in the governance failure record of 2022-2025: the Compound governance attack, the Arbitrum AIP-1 crisis, Uniswap’s years-long fee switch paralysis, and the persistent voter apathy that leaves most routine DAO governance to a handful of professional delegates. These failures are not random — they follow predictably from architectural choices made at protocol formation.

This analysis provides a systematic comparison of the principal governance models currently in use across major DAOs, their structural strengths and weaknesses, and guidance on which model is appropriate for which protocol.

Model 1: Pure Token Voting — The Original and the Problem

Pure token voting is the foundational DAO governance model: one token equals one vote, proposals that exceed a quorum threshold and achieve a majority of votes cast are executed. This is the model Uniswap launched with, the model Compound pioneered, and the model that dozens of DeFi protocols copied without modification.

How it works: Token holders delegate their voting power (to themselves or others), proposals are submitted by addresses meeting a minimum token threshold, voting runs for a defined window (typically 3-7 days), quorum and majority requirements must be met for execution, and passed proposals proceed through a timelock before auto-execution on-chain.

The structural problems with pure token voting:

Plutocracy. One token equals one vote means wealth equals power. The largest token holders — typically early investors, founding team members, and VCs — have disproportionate governance influence. This is not merely a fairness concern: it creates a systematic bias toward decisions that favour large holders over small holders, protocol users over the broader ecosystem, and short-term token price over long-term protocol health.

Rational ignorance and voter apathy. Research and voting on governance proposals has a cost (time, attention, gas). For most token holders, the expected influence of their vote — which is small relative to large holders — is insufficient to justify that cost. The rational response is non-participation. The result is chronic low participation in routine governance, with turnout figures of 3-8% being common. This low participation makes quorum thresholds difficult to reach for legitimate proposals and easy to reach for malicious ones with a motivated, concentrated attacker.

Governance attacks. As demonstrated by the Compound incident in 2024, pure token voting without safeguards is exploitable. A sufficiently capitalised attacker can acquire voting power through open market purchases, accumulate delegate agreements, and pass governance proposals that serve the attacker’s interests. The absence of any non-token-weighted check — a council, a veto mechanism, a time-extended challenge window — means that malicious proposals can be legitimised through the governance process itself.

Speed vs security tension. Governance that operates on 3-7 day voting windows cannot respond quickly to security crises. But shortening voting windows reduces deliberation time and increases the attack surface for rushed, poorly-reviewed proposals.

The verdict on pure token voting: It is appropriate as a temporary model for early-stage protocols with small, engaged communities where the assumption of meaningful participation can be sustained. It is unsuitable for mature protocols with large, distributed token holder bases and significant treasuries.

Model 2: Delegated Token Voting — The Standard Response

Delegated voting, pioneered by Compound and adopted by Uniswap, Aave, Arbitrum, and most major DeFi protocols, addresses rational ignorance through a representative model. Token holders delegate their voting power to active participants — professional delegates — who vote on their behalf.

How it works: Token holders assign their voting power (via delegation transactions) to delegate addresses. Delegates vote with the combined power of their own tokens and all delegated tokens. Token holders can revoke delegation at any time. Delegate transparency is maintained through public voting history, delegate statements, and governance forum activity records.

What delegation solves: Rational ignorance, to a significant degree. A token holder who cannot research every governance proposal can delegate to a professional governance participant who does. The delegation ecosystem creates accountability mechanisms: delegates who make poor decisions lose delegators; delegates who explain their reasoning and engage with the community gain delegators. This creates a market for governance quality.

What delegation does not solve:

Concentration. Delegated voting consistently produces a small number of very large delegates. In every major DAO with delegated governance, the top 10 delegates control 30-60% of effective voting power. This concentration occurs because: large token holders delegate to their own addresses, famous/credible delegates attract large delegations, and the economics of governance quality (one excellent delegate vs many mediocre ones) create natural consolidation.

Accountability at scale. Delegates who control hundreds of millions of dollars of voting power are not proportionately accountable. A delegate can vote incorrectly, lose some delegators, and continue to exercise decisive governance influence with their remaining delegated power. There is no equivalent to the director’s fiduciary duty that exists in corporate governance.

Competitive interests. Professional delegates often include protocol investors (VCs), competing protocols, and large DeFi firms whose interests are not perfectly aligned with all token holders. Delegate transparency (public voting records) is necessary but not sufficient to detect conflicts of interest.

The verdict on delegated token voting: The correct model for any protocol with a large, distributed token holder base. Substantially better than pure token voting for most governance purposes. Its concentration problem requires additional architectural responses — Security Councils, constitutional constraints, or multi-chamber governance.

Model 3: Security Councils — The Attack Defence Layer

The Security Council is an architectural innovation specifically designed to defend against governance attacks while preserving the legitimacy of token-holder governance. The model was pioneered by Arbitrum DAO and adopted in response to the Compound governance attack by Compound DAO.

How it works: A Security Council is a multi-signature body of elected community members with authority to take defined emergency actions without a full governance vote. Actions typically include: pausing protocol operations, reversing passed proposals during a challenge window, and executing emergency protocol upgrades to address security vulnerabilities.

Arbitrum’s Security Council model is the most mature implementation:

• 12 elected members, elected by ARB token holders
• 9-of-12 threshold for emergency actions (e.g., emergency contract pauses)
• 7-of-12 threshold for non-emergency protocol upgrades that bypass the normal governance timeline
• Six-month cohort terms with staggered elections (two cohort elections per year)
• Explicit constitutional authority defined in the Arbitrum Constitution
• Security Council authority is limited to defined action types; it cannot modify governance rules or transfer treasury funds without full DAO governance

Compound’s adapted Security Council model was adopted post-2024-attack:

• Multi-sig with veto authority over queued proposals within the timelock window
• Smaller council than Arbitrum; faster formation in response to crisis
• Authority specifically limited to blocking malicious proposals; cannot initiate protocol changes independently

What Security Councils solve: Governance attacks. The challenge window between proposal passage and execution (the timelock) combined with a Security Council mandate to monitor queued transactions and veto malicious ones creates a practical defence layer. An attacker who passes a malicious governance proposal faces a Security Council that can block execution before it occurs.

What Security Councils introduce: A centralisation vector. A Security Council with veto authority over any governance decision is a meaningful concentration of power. If the Security Council itself is captured — through social engineering, legal pressure, or conflicted interests — it can block legitimate governance decisions. The key architectural question is defining the precise scope of Security Council authority: broad enough to stop attacks, narrow enough not to become a governance oligarchy.

The verdict on Security Councils: A necessary addition to the governance architecture of any protocol with a treasury exceeding $500 million or a governance system with evidence of plutocratic concentration or attack risk. Not a governance model in itself, but a critical security layer over token-based governance.

Model 4: Optimism’s Bicameral Governance — The Structural Innovation

The Optimism Collective’s governance architecture is the most intellectually innovative governance model in DeFi — a genuine departure from the token-voting paradigm rather than an iteration of it.

The core insight: Different governance decisions require different legitimacy sources. Economic governance decisions (protocol parameter changes, treasury allocation, fee structures) should reflect economic stakeholder interests — token holders are the appropriate decision-making body. Public goods allocation decisions (retroactive public goods funding, ecosystem development grants) should reflect civic values and impact assessment — economic interests corrupt this process because large token holders will rationally favour public goods allocations that benefit their portfolio.

The Token House: OP token holders vote on protocol governance proposals through the Token House (managed on the Agora platform). Standard delegated token voting applies. The Token House governs: protocol upgrades, project incentive programmes, treasury allocations, and cross-chain governance decisions. This is the economic governance chamber.

The Citizens’ House: Non-transferable “citizenship” — awarded through attestations based on demonstrated contribution to the Ethereum/Optimism ecosystem — gives holders voting rights in the Citizens’ House. Citizens’ House governance covers: Retroactive Public Goods Funding (RetroPGF), the mechanism by which the Optimism Collective distributes grants to Ethereum public goods providers based on demonstrated impact rather than forward-looking promises. Citizenship cannot be purchased or transferred; it can only be earned through ecosystem contribution.

Why the separation matters: In a single-chamber token-weighted system, VC investors, large token holders, and protocols with large token positions dominate public goods allocation. Their economic interests will rationally influence public goods allocation toward projects that benefit their holdings. The Citizens’ House removes this conflict by giving non-transferable governance power to people who have demonstrated contribution — whose incentives align with ecosystem health rather than token price.

The optimistic approval model: Routine governance actions — those within defined parameters and consistent with established governance frameworks — are approved “optimistically” unless challenged within a defined window. Full deliberation is reserved for contested decisions. This dramatically reduces governance overhead for mature, stable protocols without sacrificing the right to challenge any action.

Limitations of Optimism’s model: It is complex. Managing two chambers with different voting mechanics, different legitimacy sources, and different quorum requirements is operationally demanding. For early-stage protocols without the governance infrastructure to manage this complexity, it is over-engineered. It is also community-specific: the Citizens’ House model depends on a large, active community of recognised contributors, which Optimism has but most protocols do not.

The verdict on Optimism’s bicameral model: The most thoughtful governance architecture in production at scale. Appropriate for mature Layer 1 and L2 protocols with large, active communities and significant public goods spending. Provides the most credible solution to the conflict of interest between economic governance and public goods allocation.

Model 5: MakerDAO/Sky Endgame — Governance Restructuring at Maximum Complexity

MakerDAO’s Endgame governance restructuring — unveiled in 2022 and progressively implemented through 2024-2025 — represents the most ambitious attempt to resolve the governance scalability problems of a large, mature DeFi protocol.

The diagnosis: After years of operation, MakerDAO governance had become dysfunctional. Core units had accumulated excessive influence. Governance proposals were too numerous, too technical, and too interrelated for the average MKR holder to evaluate meaningfully. The governance system was nominally decentralised but practically oligarchic.

The Endgame response: A radical structural redesign with several key elements.

SubDAOs. Endgame creates SubDAOs — semi-independent governance organisations with their own tokens, their own treasuries (funded from MakerDAO), and their own governance processes for their specific functional areas (Spark Protocol for lending, Sakura for RWA, others). SubDAOs handle operational governance, reducing the decision-making burden on MakerDAO’s mainnet governance.

Governance token restructuring. The MKR token is being migrated to SKY, with new token economics designed to incentivise governance participation and align long-term incentives.

Governance delegation and participation incentives. Endgame introduces delegate farming — explicit token incentives for active governance participation — addressing the rational ignorance problem through economic incentivisation rather than social pressure.

Executive votes vs polling votes. MakerDAO maintains a two-stage governance process: MIP (Maker Improvement Proposal) polling determines community preference, while executive votes execute the change on-chain. The Endgame restructuring streamlines this process for SubDAO-level decisions.

The verdict on MakerDAO Endgame: The most ambitious governance redesign in DeFi history, addressing real structural problems. Its complexity is also its principal limitation — Endgame requires sophisticated governance participants to understand its architecture. Whether SubDAO governance effectively distributes decision-making without fracturing community coherence remains to be demonstrated over time.

ENS DAO: Constitutional Constraint as Governance Architecture

ENS DAO’s governance architecture adds a dimension not present in most governance models: constitutional constraints on what governance can decide.

The ENS Constitution establishes five articles that constrain the DAO’s authority. Name owners’ rights cannot be revoked by governance. Registration fees cannot be structured for profit maximisation. ENS must operate as a public good. These constitutional constraints cannot be amended by simple majority vote — they require supermajority support and extended deliberation.

The insight: Governance legitimacy requires both the power to decide and the constraint not to decide certain things. An unconstrained DAO governance system can make any decision, including decisions that betray the protocol’s users. Constitutional constraints define the boundaries of legitimate governance authority and signal credibly to users that certain fundamental protections cannot be removed through a governance vote.

ENS’s constitutional model has been influential: several other protocols have adopted governance charters or constitutional documents that limit governance authority over fundamental user rights. This trend toward constitutional governance reflects a maturing understanding that governance is not merely about the ability to make decisions, but about the legitimacy of the decision-making process.

Choosing the Right Model: Protocol Stage Framework

No single governance model is optimal for all protocols at all stages. The appropriate governance architecture depends critically on where a protocol is in its lifecycle.

Early-stage protocol (0-12 months post-launch, small community): Pure token voting with a small, engaged community can function effectively. The community knows each other, proposals are relatively simple, and quorum can be reliably achieved. Security Council overhead may be unnecessary at this stage if the treasury is small and the community is cohesive.

Growth stage (1-3 years, expanding community, meaningful treasury): Delegated voting becomes essential as the token holder base expands beyond active governance participants. Snapshot for temperature checks plus on-chain Governor for binding votes is the standard architecture. If treasury exceeds $100 million, a minimal Security Council (or at minimum enhanced timelock monitoring) is prudent.

Mature stage (3+ years, large distributed community, billion-dollar treasury): Full institutional governance architecture required. Delegated voting plus Security Council plus governance forum plus delegate transparency infrastructure. If the protocol is Layer 1 or L2 infrastructure with public goods obligations, consider Citizens’ House or equivalent civic governance layer.

The Ethereum Foundation model — credible neutrality: For protocols that choose not to have aggressive on-chain governance (particularly Layer 1s where governance minimisation is a security property), the Ethereum Foundation’s approach — active research and resource allocation, minimal on-chain governance footprint, explicit neutrality — represents an alternative that trades governance expressiveness for governance security.

This comparative analysis is informational only. Governance architecture decisions for specific protocols should involve qualified governance designers, legal counsel, and community consultation.

Published by The Vanderbilt Portfolio AG, Zurich, Switzerland. Author: Donovan Vanderbilt.

Related Coverage

• On-Chain Governance: Token Voting, Multisig, and DAO Governance Mechanisms
• The Problems With Token-Weighted Voting
• Delegated Voting: How DAO Delegation Actually Works
• Optimistic Governance: Lower Barriers, Higher Risk?
• DAO Governance Activity Tracker: Proposals, Votes, and Participation 2025
• Compound DAO: DeFi’s Governance Pioneer and the Security Council Revolution
• Swiss Stiftung vs Cayman Foundation: The Two Dominant DAO Legal Wrappers Compared