Many cities in the developing world have implemented the driving restriction policy to address transport-related problems (e.g., traffic congestion and air pollution). However, the evidence of the effectiveness of this policy exhibits significant regional variations. Using an abrupt policy change in Chengdu, namely the tightening of the driving restriction policy by expanding the regulated area, as a natural experiment, this study provides new evidence of the effects of driving restrictions based on the difference-in-differences approach. On the one hand, we leverage the temporal variation in the number of cars being banned on roads due to the traditional superstition (number 4) in China and the spatial variation in restricted and unrestricted areas to verify the causal effect of driving restrictions on air quality. We find that policy tightening has caused a significant reduction in major air pollutants of PM2.5, PM10, CO, NO2, and SO2 and a decrease in the composite air quality index in a year after policy tightening. Furthermore, we observe that policy tightening has resulted in an increase in the demand for housing units located near subway stations in areas subject to the new driving restriction policy. This study affirms the effectiveness of driving restrictions in the absence of confounding influences from other travel demand management strategies. Accordingly, in the face of restrictions on private car use, it is important to increase the supply of public transport and affordable housing with good access to public transport.