The Inevitable Collusion in Token-Curated Registries

Collusion is the equilibrium state for any TCR with meaningful economic value. The game theory of staking tokens to vote on list membership creates a direct financial incentive for voters to coordinate and extract value from applicants, turning curation into a rent-seeking market.

The 'Dilution Defense' is a myth. Proposals like requiring voters to hold the listed token (e.g., a Curve gauge-style system) fail because large stakeholders can still collude to list inferior projects, extracting fees while diluting the cost across the entire token holder base.

Real-world evidence is consistent. Early TCR experiments like AdChain and the Kleros-curated Tokenlist demonstrated that low-stakes lists are ignored, while high-stakes lists become targets for sybil attacks and voter cartels, validating the theoretical failure mode.

Governance becomes a yield asset. In a TCR, staking tokens to curate a list creates a direct financial return. This transforms a governance action into a yield-bearing instrument, aligning voter incentives with fee extraction, not protocol health.

Collusion is the rational equilibrium. Entities like large holders or professional delegates in systems like Curve Finance or early The Graph curators maximize returns by forming voting cartels. This is not corruption; it is game-theoretic optimization of a flawed mechanism.

The price of a vote is public. Unlike opaque political lobbying, a token's market price transparently signals the cost of influence. This creates a liquid market for control, where the highest bidder deterministically steers outcomes, as seen in contentious MakerDAO executive votes.

Evidence: The fundamental dilemma of on-chain voting is that any mechanism valuing participation with a token will see that token's value converge to the price of its purchased influence, divorcing it from any underlying utility.

Quadratic voting fails in practice because it assumes costless identity creation. In reality, Sybil attacks are cheap. Projects like Gitcoin Grants demonstrate that sophisticated collusion rings, not individual voters, determine outcomes.

Token-curated registries are markets. Any market with a governance token is a vote-buying market. Rational actors form cartels to pool voting power, a dynamic seen in early TCR experiments like adChain.

Collusion is the equilibrium. The cost to bribe a quadratic voting system scales with √N, making large-scale manipulation economically viable. This isn't a bug; it's the Nash equilibrium for any financially incentivized curation game.

Evidence: Analysis of Gitcoin Grants rounds shows a small number of funding pools consistently capture the majority of matching funds, indicating coordinated Sybil campaigns that exploit the quadratic formula.

Collusion is the equilibrium state for any TCR with significant value. The economic reward for validators to form a cartel and extract value from the registry always exceeds the reward for honest, competitive curation.

Better design cannot solve this. Adding slashing, reputation, or complex voting like conviction voting only raises the collusion cost. It does not eliminate the fundamental profit motive, as seen in early experiments like AdChain and Kleros.

The data proves the failure. No major TCR operates at scale for critical infrastructure. The model was abandoned for decentralized sequencers and oracle networks like Chainlink, which use explicit, non-speculative work for rewards.