Association for Public Policy Analysis & Management

This paper provides evidence that even short-term shocks to student cognitive performance have long-lasting consequences for human capital development. We examine how exposure to crime in a student's local area affects high-stakes test scores and subsequent upper-secondary school placement in the context of Mexico City’s centralized high school admission system.

To do so, we combine administrative records on individual-level test scores with incident-level crime reports from 2013 to 2016. We exploit geographical and temporal variation in exposure to violent crime in close proximity to schools during the run up to exam dates to identify the effects of interest.

The richness of our data and context allow us to move beyond previous studies in two ways. First, using individual-level test scores and mental health outcomes, we provide evidence of poor matching between female students and preferred high-schools, along with a mechanism behind this gender disparity.

We show that exposure to violent crime in the week before a high-stakes exam lowers females’ test scores by 11 percent of a standard deviation. As a result, 19 percent of female students exposed to violent crime are subsequently assigned to less-preferred, lower-quality high schools. We find no such effect for males, leading to a further gender disparity in test scores and later education outcomes. We also show that crime-induced concentration problems are an underlying mechanism behind the detrimental effects on test scores.

Our combined results shed light on potential policy interventions for closing gender disparity in educational outcomes.