Dynamic effect of subway expansion on the spatial and temporal distribution of road congestion

Subways have long been regarded as an efficient solution for reducing road congestion, particularly during rush hours. However, less is known about such an effect’s dynamic pattern and spatial networks, especially when population density and residential locations may change in response to subway investments in the long term. This paper examines the impacts of subway investments from both temporal and spatial perspectives, using the use of high-frequency subway ridership data from Beijing during 2017–2019. The empirical analysis demonstrates significant spatial disparity in the effects of subway on nearby road congestion. Specifically, new subways in the city center decrease the congestion level near subway stations, whereas new suburban stations are more likely to increase road congestion in vicinity. As the distance from the city center increases, the congestion reduction effects of central lines and the congestion increase effects of suburban lines gradually diminish. When comparing the immediate effects (one week before and after) and lasting effect effects (one year before and after), the results suggest that this spatial pattern persists over time. Furthermore, an interesting temporal pattern emerges: the congestion increasing effect spreads more instantly for subway expansion (the effect is statistically significant in the weeks), while the congestion reduction effect emerges more gradually (the effect is statistically significant after one year). In addition, there is spatial heterogeneity in subways’ spillover effects. In city centers, they disperse traffic, reducing congestion near stations but increasing it on other stations in the line. In suburbs, they concentrate activity, raising local congestion while alleviating it elsewhere along the line. Preliminary calculations suggest that the overall time-saving benefits from the introduction of new subway stations closer to the city center are 2.53–23.67 s per vehicle. The price of housing surrounding newly built stations increased by $736–1,096 per square meter, which also suggests a latent adjustment in location preferences. In addition to validating the overall benefits of subway investments in metropolitan areas, our study further reveals the heterogeneous spatial and temporal effects of new subway lines. The dynamic and interactive effects of subway investments, road congestion, and residential changes call for integrated approaches in transport and city planning.