Uber and public transport interaction: An analysis of empirical evidence
Public transport is at the heart of most well designed urban mobility systems. In recent years, ride-hail has provided a valuable on-demand transportation layer on top. Understanding how the two interact with each other is an important policy question and vital to reducing dependence on the private car, the dominant mode of transportation in most cities. Given the high stakes for livable cities, the environment, and access to mobility, opinions on the topic can be fraught and cover the entire spectrum from “ride-hail is a boon for public transport” to “ride-hail is destroying public transport”. But what does the data say?
In a new research collaboration with the Smart Public Transport Lab at TU Delft, we study this topic by looking at empirical data on Uber trips and overlaying public transport trip data on top. We focus on three European and three American cities in October 2018 — Amsterdam, Warsaw, Stockholm, New York City, Washington, D.C., and Houston. We compare the travel time and cost of the two modes to assess the overall improvement in service accessibility across the city. In addition, we look at rider demand data to understand how riders choose between the two modes.
Our findings show that ride-hail and public transport combine to provide improved accessibility (as measured by travel times) in urban areas and that the interaction between the two alternatives depends on the relative quality of service provided. These findings are explained below.
When looking at travel times, we find that travel times for rides with Uber on average are shorter than public transport travel time. This is predominantly due to much shorter out-of-vehicle times (walk and wait times) on Uber trips. Less than 2% of the Uber trips experience an out-of-vehicle time longer than 15 minutes. The corresponding number for public transport is between 42% in Amsterdam and 65% in New York City.
By comparing travel times, we also find that between 1 out of 5 (Amsterdam) and 2 out of 5 (Houston) Uber trips have no viable public transport alternative (the fastest public transport alternative would be at least twice the travel time of the Uber trip). We can safely infer that there is no minimal competition between rides with Uber and public transport for these trips.
Unsurprisingly, rides with Uber are more expensive than public transport. The cost difference is on average larger than the time difference when converting the latter using the local value of time. In other words, on average, the travel time savings alone do not make up for the additional cost. However, all observations in our sample pertain to cases where individuals opted for the ride-hailing alternative; by only looking at Uber data, we don’t observe cases where individuals opted to take public transport instead. This choice to take a ride with Uber rather than public transport can potentially be explained by these riders having a higher than average value of time for that trip, valuing more than just travel time savings by attributing value to factors not considered in this study such as safety, reliability and comfort or never considering public transport as an alternative.
We use Uber and public transport demand data in Washington, D.C. to understand behavior in cases where a viable public transport alternative is available. We find a strong negative correlation between the travel time competitiveness of public transport and the share of rides taken with Uber. In other words, more riders choose public transport over Uber in areas and time periods where public transport travel times are more competitive. This is clearly illustrated in the map below where the relative demand for rides with Uber is low along metro corridors.
Finally, we study the added value of Uber to the service accessibility offered by public transport. We find, in general, that the Uber platform helps improve service accessibility across the entire city by offering an additional transportation option. The biggest increase in service accessibility is in areas which are not well served by urban rail.
This new research shows that the two alternatives together improve and expand accessibility in urban areas. Though riders occasionally choose between public transport and ride hail, the former is generally the superior option in areas and times when service is fast, frequent and reliable. This is consistent with recent findings in a TRB report that suggest improvements in quality of transit result in higher transit ridership.The high proportion — between 1 out of 5 and 2 out of 5 — of ride hail trips with no viable public transport alternative suggest that this option, despite higher fares, is helpful to cover the spatial and temporal gaps in public transport service. Importantly, these two modes are working together to enable a multimodal lifestyle by strengthening the basket of transportation options vigorously competing against the private car.
You can read the full working paper here
