Working Papers
[1] Visual Information in the Age of AI: Evidence from Corporate Executive Presentations (with Sean Cao, Yichen Cheng, Yusen Xia, and Baozhong Yang)
We extract visual features from corporate presentation images with large image models and find that forward-looking operational information is associated with higher short-term abnormal returns and stronger long-term operational performance (whereas other types of visual information are not). The capability to process such information contributes to an information gap between AI-equipped investors and others—a phenomenon we refer to as AI divide.
Presented: MFA 2025, NBER SI 2024 Big Data and High-Performance Computing for Financial Economics, Purdue Fintech Conference 2024, ECGI 2024, FARS 2024, HARC 2024, The 2023 Research Conference on Capital Market Research in the Era of AI, CICF 2023, 2023 Hong Kong Conference for Fintech, AI, and Big Data in Business, AAAI 2023.
Awarded: 2023 PanAgora Asset Management Richard Crowell Prize (third place)
Featured: Columbia Law Blog.
[2] The Politicization of Social Responsibility (with Todd A. Gormley and Manish Jha)
(Revise & Resubmit at Journal of Finance)
State-level politics appear to impact institutional investors’ voting decisions. Support for SRI proposals is lower in the same state when it is led by Republicans instead of Democrats.
Presented: AFA 2025, Edinburgh Corporate Finance Conference 2025, European Finance Association Annual Meeting 2025, Drexel Corporate Governance 2025, Isenberg Finance Conference 2025, Clemson ESG and Policy Research Conference 2024, NBER SI 2024 Corporate Finance, Stigler Center-CEPR Political Economy of Finance Conference 2024.
Featured: NBER, Promarket, Harvard CorpGov, VoxEU, ECGI.
[3] Heads I Win, Tails It’s Chance: Mutual Fund Performance Self-attribution
I develop a large language model (LLM) architecture to extract attribution information from mutual funds' self-assessments of performance in their shareholder reports. On average, mutual fund managers exhibit a pronounced self-attribution bias—they are 40% more likely to attribute performance successes than failures to internal factors. Funds with stronger self-attribution bias engage in excessive trading, which in turn harms their future performance.
Presented: Junior Academics Research Seminars (JARS) in Finance 2024
Featured: Morningstar, VettaFi.