The effect of violent crime on teenage pregnancy in Mexico

Abstract

This paper explores the effect of violent crime on teenage pregnancy in Mexico using data from a nationally representative longitudinal survey conducted before and after an unexpected surge in violence. In order to identify the effects, we use a triple difference-in-differences strategy exploiting variation in (i) exposure to the surge in violence between sample periods, (ii) the intensity of violence as measured by municipal homicide rates, and (iii) age of different cohorts. We find that the average increase in homicide rates over the study period reduced the probability of teenage pregnancy by approximately 1.5%. We also demonstrate that the effect is more acute among women with worse economic conditions prior to the surge in violence. Analysis of mechanisms shows that reductions in teenage pregnancy is partly explained by changes in the sexual behavior of young women.