Research
Working Papers
Platform Design in Dynamic Differentiated Goods Markets – The Case of Airbnb Job market paper
Presentations: IIOC 2026, World Congress 2025, INFORMS 2025, EARIE 2024, CESifo Warsaw Digital Workshop 2024
Abstract
This paper studies the allocation of goods in two-sided, one-to-one matching markets with finite time horizons. In these markets, the timing of seller participation distorts the allocation of goods. This comes from dynamic externalities inherent to one-to-one matching markets, where sellers fail to fully consider how their participation affects the evolution of matching on the market. Using data on Airbnb, I estimate a dynamic structural model of a two-sided matching market to quantify the externalities related to the timing of seller entry. I then evaluate policies that a platform can implement to improve total surplus on its market. Entry subsidies that encourage earlier entry increase total surplus by CHF 1.7 for every franc spent, while subsidies that encourage later entry reduce total surplus by CHF 1.0 for every franc spent. Tracing out an efficiency frontier of platform policies, I show that the platform’s current flat commission lies inside the frontier: a small subsidy for early entry, financed by a higher flat commission, raises both total surplus and the platform’s revenue.
Estimating Choice Models with Unobserved Expectations over Attributes
joint with Mathias Reynaert and Hanlin Zhao
Presentations: IIOC 2025*, APIOC Seoul 2024*
Abstract
Agents often choose based on expectations about uncertain attributes, yet researchers rarely observe the information behind those expectations. We develop two methods to estimate discrete choice models with heterogeneous, unobserved information sets. Our point-identification approach uses repeated realizations of the uncertain attribute to estimate its statistical process with observed predictors and recoverable latent factors, then represents conditional expectations as a finite mixture over these components. A complementary approach set-identifies preferences using partial information about expectations. Both methods apply to individual- and market-level data. Applications to firms’ export decisions and consumers’ automobile purchases show that our method corrects for substantial bias in preference estimates that arises when unobserved information heterogeneity is ignored.
Platforms, Spatial Frictions, and the Allocation of Drivers: Evidence from Transit Disruptions
with Kwok-Hao Lee
Presentations: SMU 2026, EEA-ESEM 2026, UEA 2026
Abstract
How does labor supply respond after sudden market shocks, and could information play a role? We study taxi responses to rail disruptions using high-frequency GPS data from Singapore. Taxi drivers serve both booking and street-hail trips. Booking requests reveal pickup locations, while street-hail relies on localized search. Taxi trips increase during disruptions, driven by bookings (10%), while the street-hail response is statistically insignificant. Booking responses remain positive farther from disrupted stations, while street-hail effects are confined close to stations. These findings suggest customer-specific information helps drivers reallocate when local search signals are noisy.
Anatomy of the Pass-through of Productivity Shocks
joint with Pablo Mileni Munari and Andrei Zaloilo
Presentations: Vigo Workshop on Dynamic Macroeconomics 2024*, Junior Search and Matching 2024*
Abstract
Workers face labor income risk when employers respond to productivity shocks by adjusting hourly wages, hours worked, and separations into unemployment. Using French matched employer-employee data, we document how employers use each margin across different jobs, and thus how earnings risk is distributed across workers. High-paying jobs adjust mainly through hourly wages, which respond to the unemployment rate 2.6 times as strongly as in low-paying jobs. Low-paying jobs instead adjust through hours and separations, with semi-elasticities 40 and 10 times those of high-paying jobs, respectively. We develop an equilibrium labor market search model with dynamic contracts in which firms share risk with their workers through these margins, choosing the least costly ones given their heterogeneous vacancy costs and their workers’ job mobility. As a result, policies that target a single margin, such as minimum wages, hours restrictions, or firing costs, can be ineffective because firms shift risk onto the others.
Selected Works in Progress
Quality, Proximity, and Price: Preschool Demand in a Near-Universal System
joint with Rachel Tan
*Presentation by co-authors.