This was created in partnership with Environment for Development .
Using individual travel diary data collected before and after the rail transit coverage expansion in urban Beijing, this paper estimates the impact of rail accessibility improvement on the usage of rail transit, automobiles, buses, walking, and bicycling, measured as percent distance traveled by each mode in an individual trip. My results indicate that the average rail transit usage significantly increased, by 98.3% for commuters residing in the zones where the distances to the nearest station decreased because of the expansion, relative to commuters in the zones where the distances did not change. I also find that auto usage significantly decreased, by 19.8%, while the impact on bus usage was small and not statistically significant. Average walking and bicycling distance (combined) increased by 11.8%, indicating that walking and bicycling are complements to urban rail transit, instead of substitutes. Furthermore, I find that estimated changes in auto usage and rail transit usage vary significantly with auto ownership and income.