Distributed ledgers and the governance of money

Blockchain technology breathes new life into the classical analysis of money as a substitute for a ledger of all past transactions. While it involves updating the ledger through a decentralized consensus on the unique truth, the robustness of the equilibrium that supports this consensus depends on who has access to the ledger and how it can be updated. To find the optimal solution, Buterin's "scalability trilemma" needs to be addressed, so that a workable balance can be found between decentralization, security (i.e. a robust consensus), and scale (the efficient volume of transactions). Using a global game analysis of an exchange economy with credit, we solve for the optimal ledger design that balances the three objectives of this trilemma. We characterize the optimal number of validators, supermajority threshold, fees and transaction size. When intertemporal incentives are strong, a centralized ledger is always optimal. Otherwise, decentralization may be optimal, and validators need to be selected from the set of users of the system.