The current study empirically tests the relationship between homicide rates and the population of indigenous Mayans at the municipal level in Guatemala. Using data from the most recent Guatemalan Census (2018) and independently aggregated national police data on homicides, we also propose models of possible pathways for the relationship between the ethnoracial composition of municipal population and homicide rates. Due to consistent non‐normal distribution of demographic data in Guatemala, we used a maximum likelihood estimation of linear regression models and nonparametric bootstrapping method to test the theoretical relationships. The results showed a strong negative relationship between Mayan majority municipalities and homicide rates, mediated by living in municipality of birth. Findings suggest that attachment to place in Mayan majority municipalities in Guatemala is a strong protective factor for exposure to homicide, even in areas of high out‐migration.