[Submitted - Last version here]

Co-author(s): Alberto Chong

Abstract: By exploiting an unexpected policy change in the form of drastic tariffs reduction across several industries in Peru during the 2000s we are able to causally show that in districts where industries’ employment are predominantly male, trade liberalization produced an increase in physical intimate partner violence of 36 percentage points with respect to control districts in our preferred specification. We find no such difference in districts where industries’ employment is predominantly female. These findings are original and consistent with several hypotheses in the social sciences. Our results are robust to falsification and placebo tests, sensitivity to initial conditions, conflation of past and current shocks, selective migration, permutation tests and input-tariffs considerations. Finally, we find considerable heterogeneity, as education and the age of first marriage appear to be key variables that correlate with our findings.