Paper Type
Short
Paper Number
PACIS2026-1485
Description
Digital platforms increasingly rely on algorithms to coordinate economic activity. In online labor markets such as ride-hailing platforms, algorithmic systems perform managerial functions including task allocation and matching between workers and customers. Prior research has largely focused on efficiency gains, paying less attention to how such systems affect workers’ capabilities. We examine whether sustained exposure to algorithmic matching reshapes these capabilities. Using trip-level data from a major ride-hailing platform in South Korea, we study drivers who transition into and out of a franchise-based automatic matching system that removes drivers’ discretion over ride acceptance. Exploiting these transitions, we estimate treatment effects using a difference-in-differences design. Our results show that exposure to automatic matching weakens drivers’ alignment with demand shocks but reduces cruising time. Drivers also become less spatially concentrated and earn higher income. These patterns persist after drivers exit the system, highlighting implications of algorithmic system design for platform operations.
Recommended Citation
Kim, Keeyoung Michael; Kang, Hyeonwoo; Lee, Dongwon; Lee, Gunwoong; Park, Jiyong; and Yoon, Seokchae, "Unintended Consequences of Algorithmic Management: Evidence from Automatic Matching Systems in Ride-Hailing Platforms" (2026). PACIS 2026 Proceedings. 7.
https://aisel.aisnet.org/pacis2026/ai_ethic/ai_ethic/7
Unintended Consequences of Algorithmic Management: Evidence from Automatic Matching Systems in Ride-Hailing Platforms
Digital platforms increasingly rely on algorithms to coordinate economic activity. In online labor markets such as ride-hailing platforms, algorithmic systems perform managerial functions including task allocation and matching between workers and customers. Prior research has largely focused on efficiency gains, paying less attention to how such systems affect workers’ capabilities. We examine whether sustained exposure to algorithmic matching reshapes these capabilities. Using trip-level data from a major ride-hailing platform in South Korea, we study drivers who transition into and out of a franchise-based automatic matching system that removes drivers’ discretion over ride acceptance. Exploiting these transitions, we estimate treatment effects using a difference-in-differences design. Our results show that exposure to automatic matching weakens drivers’ alignment with demand shocks but reduces cruising time. Drivers also become less spatially concentrated and earn higher income. These patterns persist after drivers exit the system, highlighting implications of algorithmic system design for platform operations.
Comments
03-EthicsSocietalImpact