A two-dimensional propensity score matching approach to estimating the treatment effect of urban rail transit lines on vehicle travel

We investigate the impact of the Circle Line (CL) opening in Singapore on individuals’ vehicle kilometers traveled (VKT) using the two repeated cross-sectional travel surveys. In the unmatched difference-in-differences analyses, the CL opening reduces the average VKT by individuals living in the 500-meter buffer zone (treatment) by 38.3% relative to those residing in the 500-1500-meter buffer zone (control) of CL stations. We apply a two-dimensional propensity-score-matching approach to control the spatial heterogeneity of individuals in the treatment and control zones and the temporal heterogeneity of individuals in the pre-CL and after-CL periods due to residential relocation. The CL effect increases to 48.0% with the matched samples. We apply a two-stage regression to test the impact of the CL opening on VKT in the extensive and intensive margins. We find that the CL only affects individuals’ driving decisions but does not significantly affect driving distance. The study has significant policy implications for rail transit investment decisions and sustainable urban transport.