Abstract

This dissertation examines several empirical asset pricing questions using around-the-clock high-frequency data, financial text, and AI methods. The first chapter studies systematic jump risk in equity markets by linking around-the-clock market jumps to contemporaneous news narratives identified with a reasoning large language model, and finds substantial heterogeneity in risk premia across jump types. The second chapter addresses lookahead bias in financial text analysis by developing chronologically consistent language models and showing that they perform well in both standard language tasks and asset pricing applications. The third chapter studies cryptocurrency perpetual futures, derives no-arbitrage benchmark prices under realistic payoff assumptions, and documents persistent deviations from these benchmarks and economically significant arbitrage opportunities.

Committee Chair

Asaf Manela

Committee Members

Guofu Zhou, Andreas Neuhierl; Julie Fu; Maarten Meeuwis

Degree

Doctor of Philosophy (PhD)

Author's Department

Finance

Author's School

Olin Business School

Document Type

Dissertation

Date of Award

5-5-2026

Language

English (en)

Included in

Finance Commons

Share

COinS