Ewen Gallic : ewen.gallic[at]univ-amu.fr
Avner Seror : avner.seror[at]univ-amu.fr
We estimate how the risk of domestic violence responds to male job loss, female job loss, and male unemployment benefits. Estimating this confluence of parameters on a given sample places us in a strong position to illuminate the underlying mechanisms. Our empirical analysis exploits data on the 2.4 million domestic violence cases brought to criminal courts in Brazil during 2009-2018, matched with employer-employee data covering the population of Brazilian workers, and with the social welfare register. Leveraging mass layoffs for identification, we find that both male and female job loss, independently, lead to a large and pervasive increase in domestic violence. Exploiting a discontinuity in unemployment insurance eligibility, we find that unemployment benefits do not reduce the risk of domestic violence while they are being paid, and eligible men are more likely to commit domestic violence than ineligible men once benefits expire. Our findings are consistent with job loss constituting a negative shock to income and self-esteem and a positive shock to time, and with unemployment benefits tending to offset the income shock while reinforc- ing the time shock.