Cycling infrastructure and its co-benefits: Evidence from a longitudinal natural experiment

Abstract

ABSTRACT Protected cycling infrastructure is widely promoted to encourage modal shift toward active transportation, yet causal evidence of its effectiveness on local population remains limited. This study uses a natural experiment in the Montréal region (Canada) to assess the impact of 2.5 km of new protected bicycle lanes on travel behaviour, attitudes, and neighbourhood perceptions. Longitudinal survey data were collected from intervention and control areas in 2023 (pre-intervention) and 2024 (post-intervention). We apply a difference-in-differences framework to evaluate changes to mode use, perceived barriers to cycling, and perceptions of neighbourhood conditions. The new cycling infrastructure led to a small non-significant increased in the share of weekly travel done by cycling and walking, and a statistically significant decrease in driving share. Improvements in key psychological factors influencing cycling – particularly perceptions of safety, perceived behavioural control, and neighbourhood suitability are observed. Furthermore, a reduction in perceived vehicle speed and noise and an increased in the perception of safety for pedestrians and cyclists suggest broader impacts on street dynamics. Some of these effects might stem from the removal of on-street parking. These findings suggest that even modest, low-cost cycling interventions can catalyze both perceptual and behavioural shifts over a one-year period. The study underscores the importance of evaluating not only effects of cycling infrastructure on cycling uptakes but also its broader impacts on perceptions and mobility patterns. The findings suggest the need for policies that combine short- and long-term infrastructure investments with complementary strategies to sustain and amplify impact.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 2 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.