Generalized trust is a value that leads to many positive outcomes for a society–greater tolerance of minorities, greater levels of volunteering and giving to charity, better functioning government, less corruption, more open markets, and greater economic growth.  Generalized trust is faith in people you don’t know who are likely to be different from yourself.  Yet, several people, most notably Robert Putnam, now argue that trust is lower when we are surrounded by people who are different from ourselves.  This view is mistaken.  Diversity (fractionalization) is not the culprit in lower levels of trust.  Instead, it is residential segregation–which isolates people from those who may be of a different background.  Segregation is one of the key reasons why contact with people who are different from ourselves does not lead to greater trust: Such contact may not be so frequent and it is not likely to take place frequently and in an atmosphere of equality, as argued by Allport, Forbes, and Pettigrew.  I show that residential segregation does lower generalized trust cross-nationally–and that inequality seems to have a stronger effect when segregation is high.  I also estimate a hierarchical model of trust cross-nationally and show that residential segregation drives down trust across nations.  I also show that people who live in integrated neighborhoods and who have diverse social networks are more likely to be trusting, using the Social Capital Benchmark Survey and the Knight Community Indicators Survey.  I argue that measures of diversity are surrogate measures for the shares of minorities living in a community: We know that minorities are less trusting so it is the racial and ethnic composition of a community, rather than the level of diversity, that seems to drive down trust.