Does better subway accessibility discourage driving? An analysis based on 2009 micro survey in Beijing

Beijing is investing heavily in urban rail construction, but whether it can effectively reduce congestion and related air pollution is still an open question. An empirical challenge for testing this is the fact that residential location choice and transportation choice are jointly made by a household. Using the micro data from a unique survey conducted in Beijing in 2009, we take advantage of the fanggai housing with pre-determined residential locations to address this endogeneity. Heckman two-stage regression technique is also employed to control for possible self selection bias. We find that better subway accessibility reduces a typical resident's probability of owning a car but does not affect the subsequent mileage driven. Such findings hold for the fanggai housing sample with endogenous residential location. Therefore, the development of urban rail in Beijing does discourage driving through reducing auto ownership and thus has positive congestion-mitigating and environmental consequences.