Decentralization Theater: A Three-Layer Taxonomy of Competing Forms of Centralization in DeFi

Abstract

"Decentralization" in on-chain finance has become theater: a rhetorical banner that masks the competing forms of centralization actually governing protocol behavior, producing both judicial overreach (invalidated in Van Loon v. Department of the Treasury) and regulatory paralysis (the SEC's withdrawn investigation of Uniswap Labs). Drawing on Oliver Williamson's transactioncost theory of economic governance, this Article proposes a three-layer taxonomy of DeFi as three discrete structural equilibria: Layer 1 (crypto-native, corresponding to Williamson's market), Layer 2 (hybrid, tokenized voting and delegated authority), and Layer 3 (permissionedinstitutional, corresponding to Williamson's hierarchy)-operationalized through a multidimensional coding scheme covering validator concentration, governance entropy, asset whitelisting, user permissioning, dependency profile, and legal-entity exposure. The taxonomy is validated through a triple-event study of OFAC's 2022 sanctions on Tornado Cash, the 2023 district-court affirmance, and the 2025 delisting following the Fifth Circuit's reversal, using layershare time series constructed from DeFiLlama and RWA.xyz data. Activity redistributes across layers predictably under each shock-a pattern that a binary or spectrum framework cannot produce-and the mismatch between regulatory tools and governance forms is not a failure of agency imagination but the predictable cost of asking the wrong question; replacing "is this decentralized?" with "which layer is this?" converts the current enforcement impasse into a tractable matching problem between tool and tier.