Yizhen Xie

AI and Demand-Based Option Listing

Job Market Paper

This paper shows that AI-based option listings enhance options trading volume by aligning option strikes with demand. While options trading has expanded prominently, over 80% of option listings in 2022 see zero or one contract traded per day. The inventory risk that market makers face from low-volume contracts contributes to the high bid-ask spreads of options—2.11% on average—which is an order of magnitude higher than that of stocks. To address this issue, in August 2022, Nasdaq began optimizing listed option strikes based on AI predictions of option volume. Using options transaction data from 2021 to 2023 across all options exchanges (OPRA), I find that demand-based option listings on Nasdaq increase option trading volume compared to conventional grid-based listings on other exchanges using a difference-in-differences design. I develop a model where an options exchange chooses strike listings to maximize total volume in an economy where a representative investor hedges endowment risk by trading options with strategic market makers who quote price schedules and bear inventory risk for all contracts. Consistent with the empirical findings, the model shows that AI-based option listings improve allocative efficiency by better matching strike availability with investor demand.

Conferences: AFA PhD Poster Session 2026 (scheduled); 

Inter-Finance PhD Seminar 2025; CMU Economics Lunch Seminar 2025; CMU Finance Brown Bag 2025 (scheduled)

Information Frictions and Index ETF

with Darcy Pu

We provide evidence for information frictions to explain the lack of the flow-return-fee sensitivity in index ETFs, for both sophisticated investors and unsophisticated investors. We rationalize the findings with a dynamic model with switching costs.

Presentations (*  by coauthor): LBS Transatlantic Doctoral Conference* , LBS*, CMU (2021); Inter-Finance Ph.D. Seminar, PKU Economics and Finance Alumni Forum (2022)

Structural Estimation of Executive Compensation

with Robert A. Miller, Foundations and Trends in Accounting, Forthcoming 2025

Structural estimation of executive compensation combines cross-sectional and longitudinal data relating firm performance to the wages, grants and wealth holdings of managers to quantify principal-agent models characterized by asymmetric information. The estimated models are used to measure the importance of information asymmetries, such as the degree of conflict between executives and the firms they manage, the role of human capital in mitigating conflicting interests, and the social welfare loss from moral hazard. Following a brief guide to this survey and a short review of related literature, we begin by describing the data used to estimate these models, explain the theory behind a simple static model of moral hazard, and provide a first approach to estimating them, before analyzing identification in more depth. The latter sections then show how the simplest models of moral hazard can be extended to account for other sources of hidden information and dynamic considerations that arise from the life cycle aspirations of managers.

Teaching

Principles of Microeconomics  (Undergrad)

Advanced Econometrics  (Ph.D.)

Debt Markets  (MBA)

Options  (MBA)

Real Estate  (MBA)

Global Economics  (MBA)

Macroeconomics (Ph.D.)

Econometrics  (Ph.D.)