Essays on Liquidity, Information, and Macroeconomics

Abstract

Chapter 1 proposes a theory of credit cycles driven by the private production of opaque, liquid assets (e.g., ABS or CLOs). Opacity enhances assets' liquidity, permitting greater issuance volumes, but prevents investors from determining whether the underlying projects are of low quality. Strong macroeconomic fundamentals give rise to credit booms characterized by opaque asset origination and pervasive credit misallocation. As bad projects build up in the economy, investors begin to question the value of opaque assets and eventually refuse to finance them altogether, precipitating a collapse in investment. The bust has a cleansing effect: opaque origination is abandoned, and investors no longer finance projects whose quality they cannot evaluate. I show that a policymaker would limit opaque intermediation during booms in order to moderate the subsequent bust.

Chapter 2 presents a model in which agents with heterogeneous beliefs borrow by using a physical asset and the liabilities of other agents as collateral. In equilibrium, a chain of lending emerges: each agent lends to the next-most optimistic agent and borrows from the next-most pessimistic agent. Intermediation allows optimists to lever up while pessimists invest in safe assets. In extensions of the benchmark model, I examine the implications of this arrangement for financial stability and relate the model's predictions to stylized facts.

Chapter 3, which was co-authored with Markus Brunnermeier, develops a model of digital record-keeping. Traditional centralized record-keeping systems establish a consensus based on trust in the record-keeper. Trust arises from the ability to incentivize honest reporting. Rents extracted by the record-keeper create an internal source of trust, allowing the system to be self-sufficient. Blockchains decentralize record-keeping, dispensing with the need for trust in a single entity. Some build a consensus based on externally verifiable resource costs (proof-of-work), whereas others do not (proof-of-stake). We prove a Blockchain Trilemma: it is impossible for any digital record-keeping system to simultaneously be (i) self-sufficient, (ii) rent-free, and (iii) resource-efficient. Record-keeping systems without rents or resource costs must ultimately rely on some external source of trust.