New research looks at the effect of Citi Bike docks appearing along MTA routes.
“We both had this research idea separately,” says Campbell, who recently completed a dissertation about the unintended long-term impacts of transportation. Brakewood, meanwhile, was working on a journal article that showed real-time bus information in New York led to a 2 percent increase in bus ridership. After they both queried the Metropolitan Transportation Authority for data, the agency connected them with each other. “It was a really natural collaboration,” Campbell says.
The researchers saw the before-and-after nature of Citi Bike’s rollout in a specific part of New York City as a perfect research opportunity. “It’s difficult to have this type of natural experiment,” Campbell says. “It lent itself to studying this and quantifying this in a way that we hadn’t been able to rigorously do before.”
The most common way of determining the impact of bike share on travel habits is to ask users. Surveys in Montreal, Washington, D.C., and the Twin Cities cited in the journal article show that up to 47 percent of bike-share users reported reducing their bus usage, but up to 14 percent said they ended up using the bus more. For some trips, it seems, bike share and transit are complements, while in other cases they are competitors.
Citi Bike queries its members on a regular basis, including a survey involving 645 members at the end of April, says Dani Simons, spokesperson for Citi Bike parent company Motivate. Asked how they would have made their most recent Citi Bike trip if bike share did not exist, 5.8 percent said they would have gone by bus, Simons says, while 8.3 percent would have taken a taxi or for-hire car like Uber.
Motivate does not make this survey data available to the public or to researchers, however, so Campbell and Brakewood had to find a different way to measure the impact of bike share on bus ridership.
Citi Bike posts detailed usage data on its website, as required by its contract with the city and local law. The public data allowed Campbell and Brakewood to track the location and size of each bike-share station, in addition to the number of trips from each station.
To gauge bus ridership, the researchers used detailed stats from the MTA. The data included the daily number of bus trips per route, split by the number of trips paid using the full fare or a reduced fare for seniors and the disabled.
With the data in hand, Campbell and Brakewood to set up their natural experiment, taking into account as many of the complex factors at work on New York City’s streets as they could.
To measure the impact of bike share on bus ridership, the researchers divided bus routes in Manhattan and Brooklyn into two groups: one with, and another without, a bike-share station within a quarter-mile. Some bus lines, however, only have bike share along a small portion of their route, so they were ranked in two additional ways: one measuring the number of bike-share docks within a quarter-mile, and another measuring the number of daily bike-share trips from those docks.
But Campbell and Brakewood didn’t just compare different bus routes; they also compared different types of riders on the same routes to further isolate the bike-share effect. Stats from Citi Bike and New York’s metropolitan planning organization show that people who are over age 65 or those with disabilities—two groups eligible for reduced MTA fares—are far less likely to ride bikes than the rest of the population. So it’s notable that the researchers found a drop in full-fare rides on routes near bike share, compared to the change in reduced-fare trips.
The research, which studied a period before the subway extension to Hudson Yards opened, takes into account the expansion of bike lanes and changes in bus service, like the introduction of Select Bus Service, during the study period. The model also considered other factors, such as the increasing usage of Boro Taxis and for-hire vehicles like Uber.
Finally, to make sure that the decline in bus ridership is actually correlated with the introduction of bike share, the researchers also conducted a “placebo” analysis by running the model a year before Citi Bike actually opened. Because the “placebo” did not have the same effect as the actual opening of Citi Bike in 2013, Campbell and Brakewood are confident that the introduction of bike share had a real impact on bus trips.
No matter how they sliced and diced the data, the researchers found that the introduction of bike share coincided with a reduction in bus ridership.
There was a 2.42 percent drop in bus trips associated with every thousand bike-share docks. Controlling for the expansion of bike lanes, the drop is smaller, at 1.69 percent, meaning that some bus riders could be hopping on their own bicycles instead of Citi Bike. Either way, the research finds a reduction between 12,600 and 18,100 bus trips a day after the introduction of bike share—a huge amount, considering that Citi Bike averaged a total 26,000 trips a day during the study period.
Campbell and Brakewood acknowledge that there’s room for more fine-tuning. They did not take into account changes in land use, population, economic activity, fares, and subway service. There’s also room to better calibrate for the growth of ride-hailing services, as granular data on for-hire trips has become available from the city’s taxi regulator.
They’d also like to survey New Yorkers about bike share, and learn more about how the introduction of bike share changes the transportation decisions of people people who aren’t bike-share members. “A great area for future research is to do a survey that includes both members and non-members. That would help us understand how people are getting around the city,” Campbell says. “Working with Motivate would be a wonderful way to learn more about it.”
As transit advocates and agencies increasingly focus on the basics of running fast, frequent and reliable bus service, the researchers hope to contribute one piece to solving the puzzle of falling bus ridership.
“Overall, not just in New York but also in many other major cities, we are seeing decreases in bus ridership,” Brakewood says. “Bike share is one contributor of many.”
But buses and bike share are two efficient, sustainable modes of transportation that shouldn’t be pitted against each other, Campbell adds. “This isn’t necessarily bad for public transit systems,” she says. “Instead of talking about them as competitors, the more useful conversation is about how they interact.”