Working Papers
Working Papers
Expected Option Returns and Large Language Models (Solo-Authored, Job Market Paper)
Abstract: I employ Large Language Models (LLMs), including BERT and an OpenAI model, to extract information from news articles and predict option returns. LLM-based news portfolios achieve annualized Sharpe ratios of up to 3.15 and outperform those constructed using other methods. Commonly used observable and latent factors in the stock and options markets do not explain the returns of these news portfolios. Firm-specific and pharmaceutical-related news play important roles in predicting option returns. These portfolios perform better for firms with high R&D expenditures or high stock volatility.
Presentations: ESSEC Business School Brown Bag, 2024 SoFiE Financial Econometrics Summer School
Option Mispricng and Alpha Portfolios (with Andras Fulop and Junye Li)
Reject & Resubmit in Journal of Financial Economics
Abstract: Employing a latent factor model that incorporates the time-varying dependence of systematic risk and mispricing on firm and option characteristics, we reveal economically substantial mispricing in the options market. The option alpha portfolio, constructed from individual option mispricing associated with these characteristics, yields an out-of-sample annualized Sharpe ratio of 2.70 for call options and 2.77 for put options. Commonly used observable and latent factors in both the stock and options markets fail to explain the returns of the option alpha portfolio. Risk-neutral moments, stock and option liquidity, and their interactions largely contribute to option mispricing.
Presentations: 2024 Financial Risks International Forum, 2023 Paris December Finance Meeting, 2023 Financial Econometrics meets Machine Learning, The 5th Quantitative Finance and Financial Econometrics, The 6th Asset Pricing Breakfast, ESSEC Student Research Seminar, The 19th Chinese Finance Annual Meeting
Award: Best Paper Award, 2023 Paris December Finance Meeting
Mispricing and Arbitrage Portfolios in China (with Jiawei Hong, Junye Li and Chuyu Wang)
Abstract: Relying on a latent factor model that accommodates evident structural changes and time-varying dependence of mispricing on firm characteristics, we reveal economically substantial mispricing in the Chinese stock market. For the reasonable number of latent factors equal to 4, the arbitrage portfolio constructed from estimated mispricing can earn an out-of-sample annualized Sharpe ratio of 1.79, which cannot be explained by common factor models constructed for the Chinese stock market. We find that size and book-to-market consistently contribute to both mispricing and systematic risk over time. Mispricing is much more severe in non-state-owned, high-subsidy, and small stocks than in state-owned, low-subsidy, and large stocks. We show that mispricing in China is more severe than and has low correlation with that in the US.