Empirical Behavioral Study of Airport-Serving Taxi Drivers Using Automatic Vehicle Location Data

Abstract

Many taxi drivers are willing to travel to the airport empty to pick up customers despite the much longer driving distance and waiting times. The underlying assumption of such behavior is the higher profits perceived by taxi drivers for trips originating from the airport, but lacks supporting evidence. This study validates the preceding profitability-based assumption and investigates possible affecting factors on the destination choices of vacant-taxi drivers using the automatic vehicle location (AVL) data of nearly 9,000 taxis. The analysis results reveal that airport-serving taxis earn significantly less in most time periods during the day. Nevertheless, vacant-taxi drivers are more likely to serve the airport if they have relatively higher profits in airport-originated trips. The profit and the distance to the airport have significant effects on the probability of serving the airport while the profit effect diminishes when the distance to the airport increases. Such findings are valuable to understand taxi behavior and to calibrate taxi movement models for evaluating the effects of various taxi regulation policies (e.g., price controls and subsidies or tolls for airport-serving taxis).