We test arguments that residential segregation incites racial unrest u
sing event-histories of 154 race riots in 1960 to 1993 in 55 of the la
rgest SMSAs in the United Stares. We argue that, although racial depri
vation and residential segregation may reinforce identification with r
acial boundaries and awareness of racial grievances, these problems do
not alone spark racial conflict. Instead, we find that a combination
of high levels of racial segregation followed by interracial contact g
enerates racial competition, which in turn increases the rate of ethni
c and racial unrest. Our results show that in SMSAs where residential
contact between African Americans and Whites increases, the rate of ra
ce riots increases significantly. A prior history of racial turmoil in
creases the likelihood of another race riot; the results suggest that
the rate of race riots in a metropolitan area depends nonmonotonically
on the number of previous riots in that region. These effects remain
potent in models controlling for population size, income disparity, et
hnic diversity, Black poverty rate, unemployment, and minority composi
tion.