Abstract
We study the impact of the closing of the Wilson Bridge (in the French city of Tours) to cars between August 13, 2020 and May 24, 2021 on bicycle ridership. We have daily data from bicycle counters between June 2, 2016 and May 24, 2021, as well as calendar, climate and COVID-related control variables. We argue that calculations based on average bicycle ridership comparisons tend to significantly overestimate the effect of closing the Wilson Bridge to cars (a 20% increase when comparing one specific month before and after the closing). Indeed, the increasing trend in bicycle use and growing environmental awareness are very strong confounding factors. Using an interrupted time series analysis, we find that this closure would be associated with a 7.66% (CI95%: +0.8%; +14.8%) increase in bicycle traffic. We performed several robustness tests and falsification tests that allowed us to confirm that there was indeed an increase in bicycle traffic. We believe it is important to conduct this type of study, especially when there are multiple exogenous shocks. Finally, we highlight some limitations of our methodology. In particular, because of the data used, it is impossible to clearly separate the effect attributable to a specific development from the totality of bicycle developments made during the COVID pandemic. One should stay particularly cautious with the results of the present study.