Wenxuan XU's PhD Thesis, June 13th

June 13, 2025 Research

Wenxuan XU will defend his thesis on Friday 13th June at 03:00 PM (Auditorium 6, bâtiment TSE and zoom)

Title: Essays in Industrial Organisation and Market Design

Supervisor: Professor Mathias REYNAERT

To attend the conference, please contact the secretariat of the TSE Doctoral school.

Memberships are:

  • Christian BONTEMPS – Professor in Economics, ENAC/TSE – Examinateur
  • Chiara FARRONATO, University of Harvard Business School – Rapporteure
  • Mathias REYNAERT – University of Toulouse 1 Capitole – Supervisor
  • Kevin WILLIAMS – Yale University – Rapporteur

Abstract :

This thesis comprises three chapters that explore two principal themes: (i) enhancing our understanding of decentralized two-sided matching markets and (ii) developing empirical tools to better estimate the models that inform their study. Drawing on the empirical industrial organization literature, I examine the incentive structures of two-sided markets—markets where distinct groups (such as buyers and sellers or workers and firms) interact. In such settings, individual decisions often overlook their broader market impact through spillovers, creating externalities that influence overall market dynamics. My work seeks to design market institutions that internalize these externalities, ultimately improving market outcomes.

In the first chapter of this thesis titled `Platform Design for Dynamic Differentiated Goods Markets: The Case of Airbnb', I examine the result of decentralized entry decisions of hosts to provide apartments on Airbnb, a digital platform for short-term rentals. Many emerging digital markets are organized as decentralized two-sided platforms, and their growth have raised novel questions about their design. I use detailed micro-data on the availability and characteristics of listings on Airbnb to construct a picture of the booking horizon dynamics on the platform. On the demand side, I estimate the dynamic arrival patterns of guests with heterogeneous preferences, while on the supply side, I model the forward-looking participation decisions of hosts—decisions that account for future changes in market conditions and individual opportunity costs. By employing this dynamic structural model, I quantify the extent of spillover effects and the misalignment between host incentives and the social optimum. A key finding is that hosts often enter the market too late, failing to fully internalize the value they provide to early-arriving guests by expanding their available options.

In the second chapter titled `Estimating Choice Models with Unobserved Expectations over Attributes', joint with Mathias Reynaert and Hanlin Zhao, we develop methods to estimate the information that agents use in a workhorse discrete choice setting. Consumers frequently lack perfect information about product attributes, yet the information they possess is crucial for understanding both their decision-making and to inform the design of policy interventions. Through simulation exercises and two empirical applications, we demonstrate the power of our methods and the importance of accounting for unobserved

heterogeneity in consumer information. This methodological contribution expands how information is incorporated in empirical applications and potentially provides a valuable framework for assessing the effects of information provision on counterfactual predictions in different policy settings.

The third chapter, joint with Pablo Mileni Munari and Andrei Zaloilo, titled `Anatomy of the Pass-through of Productivity Shocks', studies the insurance role of firms in frictional labor markets through the lens of dynamic contracting. Labor markets are another significant area where two-sided interactions are important. We find that firms vary in the extent to which they adjust different margins of worker earnings over the business cycle: high-paying firms predominantly modify hourly wages, while low-paying firms are more likely to adjust working hours or resort to separations. We capture these differences with a model describing how firms provide insurance to workers against productivity shocks, where firms differ in their shadow costs of adjusting various earnings margins. Our findings suggest that COVID-inspired policies, which restrict firms' ability to change wages, alter hours or manage separations, may undermine a crucial private channel of worker insurance.

Overall, this thesis contributes both conceptual insights and methodological tools to the study of decentralized two-sided markets, highlighting the importance of internalizing externalities and accounting for information heterogeneity in economic modeling and policy design.