Causal Impacts of Protected Bike Lanes on Cycling Behavior with Demographic Disparities

TL;DR

This study uses extensive bikeshare data from New York City to evaluate the causal impact of protected bike lanes on cycling ridership, revealing disparities across demographic groups and emphasizing the importance of infrastructure and targeted policies.

Contribution

It provides the first causal evidence that protected bike lanes significantly increase bikeshare ridership and highlights demographic disparities in these effects.

Findings

Protected bike lanes cause an average increase of 379 rides per station per month.

Higher effects are observed among older adults in census block groups.

Disparities exist, with less impact on communities with higher Black populations.

Abstract

Cities around the world face significant barriers to grow urban cycling, including competing budgetary priorities and car-centric streets. Thus, when making decisions regarding the installation of bicycle infrastructure, it is crucial to understand if and to what extent different bicycle-lane types increase bicycle ridership. However, associations between bicycle infrastructure and bicycle ridership have primarily been studied in the context of individual lanes and corridors, or when analyzed at the scale of entire cities, generalized across different bike-lane types. Drawing upon 72 million bikeshare trips from Citi Bike in New York, we demonstrate that there is an approximately 18% increase in bikeshare trips at adjacent stations in the 12 months following the installation of protected bike lanes (those with a physical barrier between cyclists and automobile traffic) and a 14% increase associated with painted bike lanes (where a line of pavement marking is present) and `sharrows' (where a normal traffic lane is marked with a bike stencil). However, using a difference-in-differences analysis, we detect a causal effect on bikeshare ridership only following the installation of protected bike lanes, with an average monthly increase of 379 rides per station (p<0.001). Despite this causal effect being pronounced among census block groups with higher percentages of older adults (688 rides per month per station, p<0.001), the causal effect of protected bike lanes on bikeshare ridership is absent in census block groups where the percentage of Black residents is medium to high. Taken together, these findings indicate that planners must emphasize protected bike lanes to spur ridership, and simultaneously target policies and programming to communities of color, to ensure that such infrastructure makes urban cycling a viable option for all residents.