Working papers

Squeezing More Juice Out of Lime: A Novel High-dimensional Pricing Algorithm (2024). Online Appendix

Abstract: Sophisticated pricing algorithms used by digital transportation platforms have renewed interest in price control policies, but little evidence exists on their redistributive effects. This paper studies a uniform price mandate in the market for shared electric vehicle platforms in Washington, D.C., which prohibits origin- and destination-based pricing. To compute price equilibria encompassing hundreds of prices for specific origin-destination pairs, I develop a new simulation-based pricing algorithm, adapted from the reinforcement learning literature. I apply the algorithm to a demand system estimated using geolocation data from all firms in the market. In the counterfactual exercise, I find that the redistributive effects of the price controls are mild, and mainly serve riders in the periphery of the city. Furthermore, I find that relaxing the price controls increases rides taken by consumers by 41%, firm profits by 34%, and increases consumer welfare by more than double the profit increase (80% of firm profits).

This notebook on Google colab allows you to play around with the reinforcement learning algorithm in a simple example presented in Section 2.1 of the paper.

Work in progress

Slow Auctions, Fast Prices: Optimal Design of a Trucking Procurement Platform.

Equilibria in Decentralized Freight Networks with Nick Buchholz and John Lazarev.

A reinforcement learning approach to dynamic pricing.


Rot-Jaune-Verde. Language and Favoritism: Evidence from Swiss Soccer (2023) with Alex Krumer and Michael Lechner. Kyklos, 76( 3), 380–406.