Rail Ridership Response to Planned Highway Closures

Abstract

Planned highway closures change the relative attractiveness of car and rail travel, yet large-scale revealed-preference evidence of the resulting ridership response is absent from the literature. This paper estimates the causal effect of planned highway closures on intercity rail ridership in the Netherlands. A dataset of 215 full-day closure events spanning 882 closure-days in 2024-2025 is constructed from open source loop-detector traffic intensities and cross-validated against an independent closure archive. Each closure is translated into a continuous treatment intensity per origin--destination pair using a national routing model, and effects are estimated with a fixed-effects panel design on automated fare collection data covering 2070 OD-pairs. Only closures of eight days or longer produce a significant ridership response, with a total effect of 0.375 at a doubling of car travel time, corresponding to a 3.7% ridership increase at the ridership-weighted median detour. The response is concentrated in peak periods and among passengers without an established rail habit as well as incidental passengers, and aggregates to roughly 2.73 million additional passenger-kilometres across seventeen long closures. Applied predictively, the estimates reproduce an out-of-sample 44-day closure within its confidence interval, while a single-bottleneck corridor reveals the forecasts act as lower bounds outside empirical support.