Substitutes or complements? Examining effects of urban rail transit on bus ridership using longitudinal city-level data

Despite massive investments, bus ridership has constantly declined in many cities worldwide. As public transit systems are evolving from bus-based to mixed structured, the longitudinal interactions between bus and urban rail transit (URT) have attracted the growing attention of transportation planners in many cities. Based on complex network theory and minimum spanning tree method, we examine network and scale effects of rail transit on conventional bus. Fixed effect models (FEMs) and generalized additive models (GAMs) are developed to identify non-linear causal effects of explanatory variables on bus transit ridership, rail transit ridership and total transit ridership, using China's city-level data during 2005-2020. The results showed when GDP per capita exceeded 75,000 yuan, residents would start switching away from bus; (2) for cities with advanced URT networks, simply improving bus service quality or constructing the rail network is not as effective as optimizing both together. Therefore, we recommend that for cities where the URT was less developed, financial resources need to be strategically invested in the bus transit networks so as to accommodate citizens' travel demand. Meanwhile, collaborative multimodal public transportation planning is essential for sustainable urban transportation development. The research outcomes may inform policymakers of the critical relationship between bus and rail transit and help create wise decisions on spending on public transit.