On-Chain Credit Scoring vs Overcollateralized-Only Lending

Overcollateralized-Only models, exemplified by protocols like JPEG'd and BendDAO, prioritize security and simplicity. They require borrowers to lock NFTs valued at 2-3x the loan amount, creating a robust buffer against price volatility. This model has secured billions in TVL by minimizing liquidation risk, but it severely restricts capital efficiency and excludes borrowers seeking leverage on their primary assets.

On-Chain Credit Scoring models, pioneered by protocols like Arcade and MetaStreet, introduce risk-based underwriting. They analyze wallet history—including DeFi activity, repayment records, and NFT portfolio—to assign credit tiers. This allows for higher loan-to-value (LTV) ratios, sometimes exceeding 50%, for trusted borrowers. The trade-off is increased complexity in risk modeling and reliance on the accuracy of on-chain data oracles.

The key trade-off: If your priority is maximum security, predictable yields for lenders, and a battle-tested model, choose the overcollateralized approach. If you prioritize capital efficiency, underwriting sophisticated borrowers, and expanding the addressable market, an on-chain credit scoring system is the path forward. The choice fundamentally dictates whether you are building a collateral vault or a credit marketplace.

On-Chain Credit Scoring excels at unlocking capital efficiency and expanding the accessible user base by leveraging non-financial on-chain data. By using protocols like EigenLayer, EigenCredit, or Cred Protocol to analyze transaction history, asset diversity, and governance participation, these systems can offer undercollateralized loans. This can dramatically increase Total Value Locked (TVL) utility, as seen in early pilots where loan-to-value (LTV) ratios exceeded 100% for qualified borrowers, moving beyond the strict 50-80% LTV caps of overcollateralized models.

Overcollateralized-Only Models take a fundamentally different approach by prioritizing security and simplicity, eliminating default risk through excess collateral. This strategy, foundational to protocols like MakerDAO and Aave, results in a significant trade-off: it creates a high barrier to entry and locks up capital inefficiently. However, it provides unparalleled stability, with MakerDAO's DAI maintaining its peg through multiple market cycles, backed by a robust system of liquidations and a multi-billion dollar Safety Fund.

The key architectural trade-off is between trust minimization and market expansion. Overcollateralization is a trustless primitive, requiring no subjective risk assessment. On-chain credit scoring introduces a trust layer in the scoring oracle or algorithm but enables new financial primitives.

Consider On-Chain Credit Scoring if your protocol's strategic goal is user growth and capital efficiency for a mature user base, and you can integrate with or build a reliable scoring data layer (e.g., Goldfinch for real-world assets, Spectral for NFT-fi).

Choose Overcollateralized-Only when your absolute priority is capital preservation, censorship resistance, and building a foundational, battle-tested DeFi primitive that must operate flawlessly in volatile conditions, as required for stablecoin backing or core money market protocols.