Exploring Factors Influencing the Proportion of E-hailing Taxi Trips: A Study Based on Fractional Response Model

Traditional taxi has been criticized for underserving regions far away from downtown, leading to social inequity. However, there has been no evidence so far to support this argument. With the advent of e-hailing taxi services such as Uber and Lyft, it becomes possible to test this argument. Because e-hailing taxi trips are assigned to the drivers who are willing to undertake the rides, the e-hailing taxi trips plus the traditional taxi trips can be regarded as the true travel demand served by taxi. The proportion of e-hailing taxi trips to total taxi trips is used as an indicator of to what extent area is underserved by traditional taxi. All the origins of traditional taxi and e-hailing taxi trips in New York City from January 2015 to December 2017 are aggregated at the level of taxi zones, which is defined by the New York City Taxi Limousine Commission. Fractional Response Model (FRM) is used to identify factors that influence the proportion of e-hailing trips. Results show that the proportion of African-American, the proportion of households of income less than 50,000 dollars, the proportion of household of income between 100,000 to 200,000 dollars, road density, distance to the Midtown Center have positive impact on the proportion of e-hailing taxi trips, while the population density, subway station density, bus station density, existence of airport have negative impact on the proportion of e-hailing taxi trips.