How Matching Platforms Respond to Demand Shocks: Evidence from Singapore’s Taxi Market

How does real-time information provision shape decentralized supply responses to sudden demand shocks? We study taxi responses to public transit disruptions in Singapore using high-frequency GPS data and quasi-experimental variation from rail service disruptions between 2016 and 2018. Taxi drivers serve booking and street-hail rides, which primarily differ based on the information available before a driver commits. Booking requests contain each passenger’s pickup location and destination, which are broadcast to nearby drivers before acceptance. Drivers serving street-hail trips do not have this information before the passenger enters their vehicle. We find that taxi trips rise during disruptions, but the response is concentrated in the information-rich channel. Booking trips increase by about 19%, several times larger than the street-hail response; this booking advantage is robust to controlling for local idle taxi supply and absorbing all city-wide shocks. A ride-hailing sample from a separate time period shows a similar positive response (about 13%). We interpret the booking premium as evidence that customer-specific information and the driver acceptance margin are especially valuable when demand shifts abruptly and local search signals are noisy.