Rail transit disruptions, traffic generations, and adaptations: Quasi-experimental evidence from Hong Kong

Despite a recent surge in urban rail service disruptions, rigorous impact studies are rare and the empirical literature presents mixed or highly underestimated results. In response, we examine the impacts of disrupted urban rail service on vehicle use in Hong Kong, using air quality as a proxy for the latter. We find that, on average, nitrogen oxides concentrations near an inactive metro station increased by 7.8% after the protests. This result translates into an 8.4% increase in on-road traffic intensity, given the pollution-traffic elasticity of 0.93. During rush hours, metro-station shutdowns further increased traffic intensity by ≤31.9%, suggesting an imminent need for a rail-to-road mode shift among commuters. The magnitude of the effects, however, tends to decline over time, with a 1% decline for each hour past the occurrence of a given shutdown event. This declining trend seems to reflect increased adaptation over time at both network and individual levels.