I develop a revealed-preference method for estimating neighborhood tipping points. I find that census tract tipping points have increased from 15% (1970) to 42% (2010). The corresponding MSA tipping points have also increased from 13% (1970) to 35% (2010). While tipping points are traditionally associated with the racial attitudes of white households, I find that cross-sectional differences in MSA tipping points, going from 1970-2010, depend less on differences in the racial attitudes of white households and more on the outside options faced by white households. These results support a continued role for place-based policies in mitigating residential segregation.