In a blow to the image of ride-hailing services such as Uber and Lyft as solutions to traffic congestion, researchers have found that they are actually major contributors to urban gridlock, at least in one major city, San Francisco in California, US.
These services, technically referred to as transportation network companies (TNCs), “were kind of born in San Francisco,” says Joe Castiglione of the city’s Transportation Authority.
By 2016, he adds, they had not only exploded into a global industry, but were generating nearly 200,000 rides a day in San Francisco alone.
“[In 2016] about 15% of vehicle trips in San Francisco were on TNCs,” adds Gregory Erhardt, a transportation engineer at the University of Kentucky, Lexington. To put that in perspective, fewer than 1% of the city’s vehicle trips are by taxi.
“Initially,” Castiglione explains, “there was a lot of enthusiasm because here was this great, new, affordable, fast-mobility option.” But now, he adds, “people [have] the perception that congestion in San Francisco has gotten much worse, and there were TNCs everywhere.”
To find out exactly what was going on, the city began collecting data on just when and where people were using ride-sharing services.
Since they couldn’t get the data directly from the companies, Castiglione’s team developed a computer program that pinged the Uber and Lyft apps every second for about six weeks, tallying how many cars were on the streets looking for rides at any given time.
The program, he adds, did not interfere with either company’s operations by making trip requests. Nor did it seek private user data. All it did was tabulate how many cars were out there at any given moment, as well as their locations – exactly the type of information the companies’ apps present to users seeking a ride.
“What that showed,” Castiglione says, “is that TNCs definitely concentrate in the most congested parts of [the city] and that most trips on TNCs are happening at the most congested times of day.”
To figure out how that contributed to the city’s growing gridlock, Castiglione turned to Erhardt’s team, which collected public data from vehicle navigation systems and smart phones to determine how much time people spent driving around on specific roadways, both in 2010 – before the advent of ride-hailing services – and in 2016.
From this, the researchers were able to gauge the change in overall congestion, measured by delays in travel time, everywhere in the city.
The emergence and growth of ride-hailing services, of course, wasn’t the only thing that changed from 2010 to 2016.
“The population was growing,” Erhardt says. “Employment has grown by quite a bit. Roadways have changed.”
But all of those factors, he adds, could be accounted for with a computer program he describes as “sort of like a SimCity model”.
And while that model predicted that San Francisco traffic snarls were destined to get worse, no matter what, traditional growth factors only accounted for a fraction of the gridlock increase that actually occurred.
Instead, Erhardt found that more than 60% of the congestion change was due to TNCs.
Why network company activity increases congestion remains an open question, but there are several possible factors, Erhardt and Castiglione say.
The traditional view of TNCs is that ride-hailing encourages fewer people to own cars, might contribute to carpooling, and might increase the use of public transportation by allowing people to hail a ride for the short trips to and from the nearest transportation hub, using public transport for the rest.
But the researchers found that none of this is actually occurring, at least in San Francisco. There was very little change in car ownership during the study period, public transportation ridership was flat or declining, and there was no marked increase in carpooling.
Instead, they say, about two-thirds of ride-hailing trips involved passengers who would not otherwise have been on the road. In fact, the average distance of a ride-hailing trip in San Francisco, Erhardt says, is only four kilometres, meaning that TNCs are directly competing with bus trips (which are used on average for journeys of three kilometres), bicycling and walking.
Furthermore, there is nowhere for the Uber or Lyft cars to park between rides. Instead, they have to cruise around, idle, in a process known as deadheading.
Elliot Martin, a research and development engineer at the University of California, Berkeley, who was not part of the study team, says the effect is as though each out-of-use vehicle is, in effect, “storing itself on the road while it’s waiting for its next ride”.
Alejandro Henao, a mobility researcher with the National Renewable Energy Laboratory, in Golden, Colorado, who was also not part of the study team, compares it to what happens when private drivers circle block after block, searching for a parking spot.
“[R]ide-hailing deadheading is worrisome,” he says, “as we know how much ride-hailing is growing.”
Other researchers are impressed by the new findings. “These authors did an incredibly thorough job,” says Martin.
“It’s an excellent study – quite carefully done and ground-breaking,” adds Bruce Schaller, a transportation policy consultant who has studied the effects of TNCs in New York City.
“It adds to what we know in New York City in documenting traffic impacts of TNC growth and provides clear rationale for limiting their numbers in congested urban centres,” he says.
Not that the researchers are advocating drastic restrictions on TNCs.
Henao – who collected data for a PhD thesis on the subject by the simple expedient of working as a driver for the two best known companies – notes that TNCs have a number of benefits. For example, they can do a lot to reduce parking demand at airports and other major urban destinations.
In addition, they do much to improve the lives of passengers with disabilities, who cannot easily travel using public transportation. They might also improve the environment by increasing the fraction of energy-efficient vehicles and electric cars.
In the urban core, Castiglione notes, there are ways to reduce TNC-related congestion without banishing them entirely. One is the imposition of congestion fees on to cars of all types, regardless of whether they are ride-sharers.
Another is a reevaluation of how the City allocates “curb space” — something that might allow ride-hailing vehicles to park and wait for rides, rather than deadheading.
Meanwhile, Martin says, the businesses have only been around for a few years, and we are still trying to understand how best to use them.
“That’s part of the messy process of how these transportation systems evolve,” he says. “They have an impact on people’s behaviour, that has a whole host of changes that are important for us to understand.”
Which, Castiglione says, is exactly the point of his and Erhardt’s work. “Good public policy is based on data and real information and analysis. Our motivation was to understand what is really going on.”
This content was originally published here.