Researchers have long suspected that the built environment directly affects people’s travel decisions, but this intuition, while backed by a lot of suggestive evidence, has always been haunted by the fact that people choose where they live. Our research, by exploiting the random assignment of a broad variety of people into housing units and neighborhoods, gets us much closer to seeing that the built environment really does cause people to change their travel behavior. Even within San Francisco, a more walkable, bikeable, and transit-accessible city than most other places in the U.S., transit accessibility substantially affects car ownership and travel behavior. This suggests that even more substantial household responses to increased bus frequency, for example, might be expected in places where transit service is minimal at present.