Abstract
This paper examines decentralized finance (DeFi) as an experiment in institutional design, asking whether a financial system built on programmable code can perform the functions of traditional banking without reproducing its safeguards. It begins with the technical foundations of blockchain and the two dominant consensus mechanisms, Proof of Work and Proof of Stake, showing how each anchors network security to a different economic cost and carries distinct implications for wealth and governance concentration. It then traces the shift from Bitcoin's deliberately constrained scripting to Ethereum's smart contracts, which made it possible to rebuild lending, exchange, and risk management as composable and disintermediated protocols.
The analytical core compares DeFi with the commercial bank. Drawing on Diamond and Dybvig's model of bank runs, the paper reconstructs the four institutional architectures that traditional finance developed in response to recurrent crises, namely deposit insurance, the lender of last resort, capital requirements, and discretionary prudential supervision. It argues that DeFi replicates the intermediation functions of banking while implementing none of these crisis-management frameworks, and that the determinism of smart contracts operates as both a strength and a structural constraint. The closing discussion frames the resulting tension between rule of law and rule of code, and considers the conditions under which the two systems might coexist.
This article is an English translation of a piece originally written in Italian by Cosimo Petralia for the Rome-based think tank Trinità dei Monti, where it was first published. The original version can be read at trinitamonti.org.
