Causal, spatiotemporal impacts of transit investments: Exploring spatial heterogeneity from announcement through long-run operation. (July 2022)

• Record Type: Journal Article
• Authors: Schmidt, Adam; Bardaka, Eleni; Thill, Jean-Claude
• Is Part Of: Transportation research. Volume 161(2022)
• Page Start: 151
• Page End: 169
• Publication Date: 2022-07
• Subjects: Difference-in-differences -- Spatiotemporal analysis -- Transit -- Causal impacts -- Quasi-experiment -- Anticipation effects; Transportation -- Research -- Periodicals; 388.011
• DOI: 10.1016/j.tra.2022.04.014
• Languages: English
• ISSNs: 0965-8564

Abstract

Abstract: As the successful design and implementation of alternative mechanisms for funding transit infrastructure, such as value-capture schemes, are becoming more critical, identifying the timing, duration, and spatial extent of the capitalization of accessibility benefits for nearby communities is becoming more and more important. This research makes a significant contribution to the analysis of the spatiotemporal impacts of transit systems by quantifying variations in the spatial distribution of causal effects from project announcement to long-run operation. The study develops a quasi-experimental framework based on advanced difference-in-differences specifications that enable us to capture the distribution of average treatment effects in space and time. The methodology is applied to the light rail system in Charlotte, NC, which includes an original line and its extension. A dataset comprised of the single-family house sales from the last thirty years is compiled for the study area, which contains neighborhoods in the vicinity of the light rail and two comparison areas. The estimated impacts demonstrate significant heterogeneity, on multiple facets. We find that although the shape of the spatial distribution of effects changes over time experiencing increasing curvature, the results consistently indicate highest positive impacts for properties located within 0.25 and 0.5 miles of a transit station. Differential effects are also identified between the original light rail line and its extension, such as lack of anticipation effects for the line extension. Although these results are specific to a single locale, the methods demonstrated through this study can be applied to other areas and transit systems of similar scale.