Centralized Finance and Decentralized Finance

Leili Pour Rostami, University of South Carolina

Abstract

This dissertation explores different facets of financial markets, focusing on the interplay between decentralized finance, traditional banking systems, and private market relationship lending. By examining these diverse areas, we aim to provide a comprehensive understanding of how emerging financial instruments and market behaviors interact with traditional financial systems. Decentralized Finance (DeFi) offers an escape from the challenges of centralized finance (CeFi). The first essay explores whether and how returns on DeFi instruments, particularly cryptocurrencies, manage to evade the traditional “rules” in response to exogenous central bank monetary policy shocks. The findings reveal that crypto abnormal returns decline significantly with shocks, with losses amplified by two indicators: “Age” and “Volatility.” However, certain crypto categories, such as ESG, Fintech, NFTs, Solana, and Bitcoin ecosystems, demonstrate resilience or growth, countering the overall negative trend. Additionally, negative sentiment from traditional financial institutions regarding monetary policy can result in positive crypto returns, indicating a potential shift to cryptocurrencies when traditional assets falter. This suggests that while DeFi aligns with some traditional rules, it continues to serve as a haven from centralization. In the second essay, the research shows that households’ demand for deposits declines when cryptocurrency returns are high, which in turn reduces bank lending and employment. This phenomenon is identified through the heterogeneity in banks’ geographic exposures to households with cryptocurrency investments. Banks exposed to crypto-investing households experience deposit outflows when cryptocurrency returns are high and subsequently reduce lending in those and other markets. These findings highlight a novel indirect exposure of banks to cryptocurrency markets and illustrate how cryptocurrency demand can have real effects by weakening traditional financial intermediation. In the third essay, this dissertation addresses private market questions concerning the bright- versus dark-side views of relationship lending in a dynamic setting using public market data for syndicated loans. The analysis reveals a bright-side empirical dominance for past term loan relationships and a dark-side prevalence for credit line-based relationships. These results, while novel, are consistent with predictions from prior relationship lending research. The study uses bid-ask spreads on market prices of syndicated loans, informed by lead banks acting as market makers on their relationship and non-relationship loans. Employing strong fixed-effects identification and high statistical power, this research yields significant implications for both research and policy.