Bike-sharing systems have swept the world, making them the transportation mode of choice for many urban millennials. Cities like Hangzhou and Shanghai in China and London, Paris, and New York have the biggest networks. Washington, D.C. and Chicago are among the largest in the U.S.
But anyone who has used bike share programs has run into some problems—big distances between bike stations or too many bikes at remote locations while there are too few at busy ones.
How to fix this and distribute bikes efficiently? That was what Poets&Quants’ Professor of the Week, Karan Girotra of the Cornell SC Johnson School of Business, tried to do in a new paper that both analyzes the problem and proposes solutions. The paper, “Bike-Share Systems: Accessibility and Availability,” was co-authored by Ashish Kabra of the Robert H. Smith School of Business at the University of Maryland and Elena Belavina of Cornell Johnson. It has just been published online in the peer-reviewed journal, Management Science.
PROFESSOR OF THE WEEK: KARAN GIROTRA OF CORNELL
The 38-year-old Girotra and his co-authors obtained data from the Vélib’ bike-share system in Paris, the largest such system outside China. They observed 946 bike stations in central Paris for four months, covering roughly 17,000 bicycles making more than 4.35 million trips, and then matched their findings with broader demographic and urban density data from French government statistics.
They found that “a user’s likelihood of using bike-share decreases with the distances they must walk.” In other words, bike share users are ready to make the extra effort to ride a bike long distances rather than drive or take public transport, but they’re unwilling to walk too far to find a bike to ride.
Girotra and his colleagues even quantified how far people were willing to walk to a bike station: 300 meters—about 1,000 feet, less than 1/5 of a mile, or roughly four city blocks. That doesn’t sound so far, but at 300 meters, the researchers found, potential users are “almost 60% less likely to use a station than one that originates close to the station.” And after the first 300 meters, “every additional meter of walking decreases this likelihood by 1.307%.”
BIKE-SHARE DEMAND IS VERY LOCAL
When potential users are 500 meters—a little more than six city blocks—or more away from a bike station, they “are highly unlikely to use the system,” Girotra and his co-authors observe.
“Overall, these estimates highlight the very local nature of bike-share demand and consequently the need for a very dense network of stations to make bike-share systems effective,” they find.
They also show that “increasing bike-availability increases ridership. A 10% increase in bike-availability at stations would increase system-use by about 12.211%…” Where should those stations go? “Locations where there are many points of interest and locations with lower bike availabilities,” Girotra and his co-authors recommend. “Among these, adding stations in areas closer to supermarkets provides more beneﬁts than adding them closer to public transit and other points of interest.” Just don’t forget to bring your canvas bag when you haul groceries home with you on the bike.
Girotra is the Charles H. Dyson Family Professor of Management at Johnson and Cornell Tech, a partnership between Cornell and Technion-Israel Institute of Technology whose campus is located on Roosevelt Island in New York City. His research and teaching interests range from the traditional—data analytics, microeconomics, and game theory—to cutting-edge 21st Century issues like online retail, carbon markets, electric vehicles, sustainability, and global supply chains.
He got his bachelor’s degree from the Indian Institute of Technology—Delhi and his masters and PhD in managerial sciences and applied economics from The Wharton School of the University of Pennsylvania. Girotra was one of the founders of Terrapass Inc., which has helped businesses and consumers reduce over 100 million tons of carbon dioxide emissions. He has been at Johnson since 2017, having taught at Insead for ten years immediately after he got his PhD in 2007.