| Abstract: |
Vehicle sharing systems, such as those for bicycles, scooters, and cars, are fundamental to serve transportation needs. Companies that operate these systems set prices (or fares) using algorithms to determine how much a user must pay and display the fares through mobile applications. This may result in users from different locations paying different prices for a vehicle. Moreover, the overall accessibility of these systems may be very different depending on the user’s location. Platforms and regulatory bodies should consider how to set prices fairly to minimize the inequalities experienced by users across locations, for which we provide a framework and insights in this paper. We consider two notions of fairness corresponding to price and access. Price fairness is a measure of how similar the prices at two locations are to one another. Access fairness compares the proportion of demand at each location that has access to the system, where access is a product of affordability and vehicle availability. We analyze the impact of imposing these fairness measures on the revenue of the platform, consumer surplus, and social welfare using a stylized model with two locations. Under price fairness, we analytically identify regimes where consumer surplus at both locations can increase. We show that imposing access fairness always results in a decrease in consumer surplus at both locations; although access fairness appears well intended, it actually results in all parties being worse off. To address the more general network pricing problem, we employ a convex relaxation technique that is asymptotically optimal with access fairness constraints. Lastly, we conduct a case study using real-world data of a vehicle sharing system operating in New York City. We analyze the tradeoff between fairness and various metrics, including revenue, consumer surplus, and social welfare, based on the fairness criteria we define.Funding: This work was supported by the Division of Civil, Mechanical and Manufacturing Innovation [Grant 1944428] and the Division of Information and Intelligent Systems [Grant 2147361].Supplemental Material: All supplemental materials, including the code, data, and files required to reproduce the results, are available at https://doi.org/10.1287/opre.2024.1043. [ABSTRACT FROM AUTHOR] |