The allegedly complex relationship between trust and victimization has rarely been modeled and, when done, the
effect of trust on victimization has been found not statistically significant. This study finds otherwise, estimating
an instrumental model with community data from Cali, Colombia. Cali’s dismal levels of victimization were only
second to Medellin, the most violent city of the world in the 1990s. But Cali also pioneered a strategy of social capital
formation as the backbone of a deliberate public policy to crack down on high levels of crime. This article first develops
a model of victimization that includes interpersonal trust as determinant and then instruments interpersonal trust
with district-level average trust. We argue that an individual-specific level of trust in his or her community members
does not affect the community level of interpersonal trust in the margin. However, the levels – or perceived levels – of
interpersonal trust in the community may affect the specific level of trust of an individual in other members of that
community, along with personal characteristics and experience. Using GMM estimates, this study finds evidence of a
relationship between interpersonal trust and victimization, statistically significant and negative in sign. The result is
robust across specifications of trust, victimization, and estimating techniques. We conclude that increasing trust in
trusting communities contributes to reducing victimization in its own right, although the effect is modest. Consequently,
strengthening interpersonal trust is another bullet to combat victimization but it is not a silver bullet.