MIT OpenCourseWareLecture 1: Introduction to 14.129 Blockchain and Design of Financial Systems
MIT OpenCourseWare
Course overview and framing 0:00
Robert Townsend opens by explaining that 14.129, officially titled Blockchain and the Design of Financial Systems, blends computer science and economics. The computer science side brings cryptocurrency, blockchain, tokenization, platforms, and computational algorithms, while the economics side brings contract theory, mechanism design, general equilibrium theory, and monetary theory. The aim is to examine distributed ledgers, smart contracts, and encryption, understand their assumptions and shortcomings, and explore their impact on existing financial systems and regulation. He stresses avoiding both hype and excessive criticism of technologies like Bitcoin, instead taking a balanced, objective view of how different technologies can implement the role economists call the central planner.
Blockchains and financial accounts as databases 6:03
Lecture 2 treats both blockchains and financial accounts as databases of transactions, each linked to a notion of money, whether Bitcoin-style cryptocurrency or fiat money and bank deposits. Townsend explains the blockchain's hashing structure, where data are hashed pairwise repeatedly to form blocks and chains, and notes that both systems can track multiple goods, assets, or IOUs, not just money, with money defined as whatever circulates with high velocity. He raises the tension between holding money as a store of value versus using it in transactions, contrasts this with Pacioli's double-entry bookkeeping, which treated money as a community liability rather than an individual asset, and mentions the CAP theorem, an impossibility result showing you cannot have consistency, availability, and partition tolerance simultaneously, a tradeoff he says echoes choices seen in real economies.
Readings and research-based course structure 13:33
Townsend lists optional readings tied to each lecture, including a BIS piece by Agustin Carstens on the future monetary system, a piece on a babysitting coupon economy reviewed by Paul Krugman, work by Goldstein and co-authors on payments and financial fragility, and an article by De Meijer on cryptocurrency use in Kenya. He clarifies there is no midterm or exam, and that the course is built around research, with these readings offered as possible starting points for projects rather than mandatory material.
Fragmented markets and distributed ledgers 15:32
Lecture 3 looks at fragmented markets, using the Pareto optimality criterion, a standard where no one can be made better off without harming someone else, to judge efficiency. Townsend contrasts the US regulatory approach to fragmented stock markets with a blockchain approach that accepts fragmentation but tries to reach efficient outcomes despite it. He argues decentralized solutions, whether relying only on bilateral trading histories or on carrying cash, either fail to reach the efficient Walrasian outcome or require excessive liquidity, and notes that real-time gross settlement systems similarly demand large liquidity, prompting the search for liquidity-saving mechanisms and alternatives like a central warehouse trader or overdraft accounts, each with its own drawbacks such as market power or default.
Smart contracts and coordination crises 30:00
Lecture 4 turns to smart contracts as a way to solve coordination problems among privately issued monies circulating across different locations and dates. Townsend describes how agents trading pairwise can issue and pass along IOUs, which works well until agents misjudge conditions elsewhere, causing crashes, a pattern he links to Walter Bagehot's account of bills of exchange in historical London and to modern concerns in low and middle income countries and in decentralized finance markets. He references readings on private money production without banks, the real bills doctrine versus quantity theory debate on inflation, and suggests multi-agent smart contracts as the emerging solution to these coordination failures.
Tokenized assets and settlement problems 36:31
Lecture 5 covers tokenized, programmable assets and atomic settlement, contrasting them with legacy systems that suffer from trade fails, sometimes reaching into the trillions of dollars in value. Townsend explains that tokenization can reveal private information about holdings, that systems built this way may only be partially interoperable, and that trust issues arise when a pre-existing asset must be escrowed to be tokenized. He notes ongoing efforts by central banks and international agencies, including a proposed blockchain for foreign exchange transactions and BIS proposals for unified ledgers, while cautioning about preserving what he calls the coherence guarantee of multilateral smart contracts.
Key players in risk-sharing networks 43:00
The lecture identifies which participants in a financial network are most valuable to support with liquidity. The most valued node turns out to be a trader who takes part in market clusters even when few other people are participating, especially when the cluster faces correlated shocks, low average incomes, and high risk aversion among agents. This idea can be tested by imagining a bond that pays off based on whether a specific individual is present in the market, and similar patterns show up in village studies, where key players who trade even in thin markets end up with higher average consumption, as if earning an informal premium.
Risk-sharing versus contagion views 44:31
A separate strand of research also studies financial networks and interconnected nodes, but frames the problem as contagion, where a shock spreads like a disease from trader to trader. This view favors limiting the size and connectivity of markets, the opposite of the risk-sharing perspective, which favors adding more traders to spread risk. Two references are given: a book called Contagion, which reflects how policymakers currently think about systemic risk and about containing damage, and a review piece by Martin Sumner on financial contagion.
Incentives, trust, and Byzantine generals 46:31
Mechanism design deals with getting people to report truthfully and act appropriately, illustrated by an insurance company that cannot directly observe a client's private loss. Smart contracts can replace a central planner: code is validated in advance, payouts sit in escrow, and claims are recorded on a ledger, without needing every layer of blockchain technology. Computer scientists instead think in terms of parallel computing failures and Byzantine fault tolerance, using tools like Bitcoin's proof of work to make rogue behavior statistically unlikely. The Byzantine generals problem, where two generals must coordinate an attack despite unreliable messages, shows how a workable protocol can still be undermined once economic incentives and Bayesian reasoning are introduced, a topic taken up further in lecture 8.
Encryption as the third technology leg 52:04
After distributed ledgers and smart contracts, the course turns to encryption, tracing the idea of secrecy back to Mesopotamian clay tokens sealed in envelopes to prevent tampering with shipments. Modern private and public key systems let people send secret messages, authenticate senders, and verify that messages have not been altered, using hashes, cryptographic puzzles, fully homomorphic encryption, and multiparty computation, including zero-knowledge proofs. These tools already appear in real implementations such as Chainlink, Layer Zero, and the Dfinity internet computer, adopted by industry and central banks.
Encrypted auctions and matching 56:00
Lecture 10 applies encryption to auctions, allowing bids to be encrypted while still determining a winner without a third-party auctioneer who could cheat. The same approach handles insurance cases where agents do not want to reveal their true balance sheet shocks, and it extends to centralized matching where supply and demand schedules stay encrypted yet still yield the correct equilibrium, including work MIT is doing with Visa on blockchain encryption.
Market games and decentralized exchange 57:05
The final lecture pairs Dubey's classic paper on price-quantity strategic market games, which reaches a Walrasian outcome, with SPEEDEX, a decentralized exchange sharing similarities with that model. Because finding a Nash equilibrium among many strategic agents is computationally hard, ongoing work, partly based on Costas Daskalakis, lets an algorithm compute a candidate equilibrium and signal agents their strategies directly, removing the need for endless guessing about others' behavior.
AI-generated summary. It can be wrong or incomplete - check anything that matters against the original.
