LAUSR

Abstract How to evaluate the comprehensive costs and benefits of urban light trains is a critical issue. The difference-in-differences technique is widely used in the literature. However, the conventional approach assumes the project does not affect the control group, which might be violated in real life due to social and economic interactions. We refine the DID approach and relax the no-spillover assumption. Through an example of impacts of a new light rail transit on house prices in Edmonton, Canada, we illustrate how the approach can be used to calculate three different treatment effects. The results indicate that the train system has negative impacts on values of single-detached houses nearby the train stations, and the negative effects also spill over to properties outside the treatment zone. Our findings show that the direction and magnitude of project impacts depend on the city size, the design of the train system, and relevant neighborhood characteristics.