“School starting age and juvenile crime in Chile: A Regression Discontinuity Design approach”

Abstract

This paper investigates the causal effect of school starting age (SSA) on juvenile crime in Chile. Using Chilean administrative data and exploiting the variation induced by minimum age requirements, we implement a fuzzy regression discontinuity design to estimate the effect of SSA on juvenile crime. We find that young females starting first grade at an older age have a higher propensity to commit juvenile crime. This find goes in the opposite direction one would expect based on the positive effects of later SSA found in the existing literature. This effect appears to be driven mostly by severe crimes and crimes with economic motivations. In contrast, we find no effect of SSA on juvenile crime for young males. Regarding the mechanisms driving these results, we find no effect of later SSA on the probability of dropping out of school, but we find a positive effect of later SSA on the probability of truancy for young females. Finally, we propose teenage pregnancy as an alternative explanation for our findings.