Research
Working Papers
Contextual Bias in LLM Financial Advice
(Solo, subsumes my Job Market Paper “Memory and GAI”)
Irrelevant context reshapes how language models translate financial beliefs into investment actions: they can estimate the right probabilities yet still choose the wrong asset.
--- (Scheduled) presentations at Harvard, SAIF, Fudan SOM, Fudan SIMIS, Tongji SEM, ABFER 2026, ABFER 2025 Poster session, CICF2025, CFRC 2025, AEA2026, AFA 2026 Poster session.
AI as decision maker: Ethics and Risk Preferences of LLMs (with Shumiao Ouyang and Hayong Yun)
Language models do exhibit consistent risk preferences and ethical fine-tuning has unintended consequences on their risk preferences.
--- (Scheduled) presentations at ABFER-JFDS 2024, OxNLP, GPI, Oxford Finance, SBS Board, SAIF*, PKU NSD, ZJU Econ, SFS Cavalcade 2024, CREDIT 2024, Adam Smith Junior 2024, CBF 2024, 4th Hongkong AI Conference 2025*, ES World Congress 2025, NAWM Econometric Society 2026, Paris December Meeting 2025, GSU AI conference 2025, Luohan Academy, Shanghai AI Lab, USC, CEIBS, Edinburgh, Durham CBID, Glasgow Adam Smith, Association for Computational Linguistics (ACL) 2025 BoF, Xueshuo, and OMI 2025, Fudan SOE*, CICF 2026*. (* denotes presentation by Xingjian)
Beta for Alpha: Neural Engagement in High-Stakes Financial Trading (with Liang Chen, Tse-Chun Lin, Fei Wu and Eric Zou)
Brainwaves provide a direct measure of investors' cognitive attention. Engagement level does not predict trading profit, whereas changes in engagement level do.
--- (Scheduled) presentations at 20th International Behavioural Finance Conference, CICF 2026*, CFRC 2026*, SZU, ZJU, HKU, NChengchi U. (* denotes presentation by Xingjian)
The Cross-Section of Price Changes in Prediction Markets (with Shuyang Hou and Xinwei Li) New!!
Prediction-market prices are systematically predictable: low-market-value contracts subsequently appreciate relative to high-market-value contracts, and recent price changes strongly reverse.
Publication
Clustered by images: Convolutional neural networks, investor heterogeneity, and Chinese stock market predictability
(《图以类聚：卷积神经网络，投资者异质性与中国股市可预测性》 in Chinese, with Yuqiao Fang and Feng Li)
We propose a simple but effective method to improve the predictability power of CNN in the Chinese Stock market.
--- Presentations at CFAC2024. Accpected at Journal of Management Sciences in China （《管理科学学报》, top tier journal in China）
Work in Progress
Geopolitics at work: U.S. Blacklisting of Chinese Firms and Employee Health (with Tse-Chun Lin, Fei Wu, Rong Yu, Bohui Zhang and Eric Zou)
U.S. blacklisting sparks innovation at Chinese firms—but workers bear the cost in greater job pressure and poorer health.
--- Partnered with Health-100 (Meinian Onehealth Healthcr Hldngs/美年大健康).
Improving Investor Trading Habit via Digital Companion (with Xiaomeng Lu, Jun Qian and Shang-Jin Wei)
We conducted experiments and found that digital nudging is effective at changing investors' behavior.
--- Partnered with the Ant Group.
Inactive Working Papers
Carbon emission and asset prices: new evidence from machine learning (with Feng Li)
We predict the carbon emissions of US-listed firms with XGBoost and find a reversed carbon premium after 2016. (This is my second-year summer paper.)
--- Presentations at CICF2023, CFRI-CIRF 2023 joint conference, CMCSR2023, SBSICF2023