The short-term effect of quarantine requirement on domestic violence

Quarantines implemented worldwide following the spread of the coronavirus resulted in many people having to spend more time with their families and partners, which can potentially lead to an increase in intra-family violence. Our study seeks to assess the extent to which this policy response to the coronavirus crisis had unintended consequences in the level of domestic violence. The goal of the project is to inform policy makers about the costs of social distancing measures, the characteristics of the affected population, and the need to complement quarantines with targeted services to curb intra-family conflict and violence. Assessing the impact of coronavirus quarantines on domestic violence has, of course, a series of challenges. Although most governments soon recognised restricting movement of people was important to limit the spread of the disease, their reaction varied enormously in terms of the type of restriction they imposed, from localise suggestions to full national lockdown. It is impossible to identify the effect of quarantines when they stem from recommendations, and only in a subset of countries did the government impose compulsory quarantines. Restrictions to movement also varied substantially in their timing with respect to the advance of the corona. Many, especially developed countries in the Northern Hemisphere, acted only when the spread of the disease was already imposing a serious threat to the population, making it hard to figure out whether domestic violence is consequence of the confinement per se, or partly consequence of the stress triggered by the fear of the disease. In this respect, Argentina makes it an ideal scenario to carry out this study. The disease arrived relatively late to the country and, since by then there was already a consensus that the virus represented a serious threat, the Argentine government reacted fast and strong, with a full national lockdown. In the way this lockdown was implemented, however, it exempted from the obligatory quarantine a very specific subset of the population, providing exogenous variation to the people within the country that were effectively under compulsory confinement. The first recorded case of coronavirus in Argentina appeared on March 3, more than two months after the beginning of the outbreak in China, and the first death on March 7. The WHO declared the corona outbreak a pandemic on March 11, more or less when first ‘autochthonous’ cases begin to surface in the country. Despite these low figures, with the world in alert, the Argentinean government decided to take extreme measures to control the disease. Following the cancelation of a series of mass events and the closure of schools and universities on March 16, a full compulsory national lockdown was declared on March 20, when a total of only 3 deaths were confirmed within the country (Argentina has a population of 45 million). This marks a stark contrast with most other places in the world. Asian countries had mostly localised lockdowns or provided ‘recommendations’ at national level. The US and Canada also reacted late and with localised policies, many of them rather timid. Most European countries did eventually engage in national compulsory lockdowns, but only when they were already quite compromised in terms of cases and deaths. In Argentina the lockdown effectively affected the whole country when the effect of the actual disease was marginal. Twenty days into the lockdown, the total death toll is 71 (less than 2 death per million population). More than 600 had died in Italy and nearly 200 in Spain when national lockdowns were declared, and twenty days into the lockdowns they had each more than 10,000 deaths (more than 300 death per million population), nearly 150,000 recorded cases, and arguably many more unrecorded. Two features, then, make Argentina a particularly interesting case to study domestic violence under the quarantine. First, because the full lockdown took place in an environment where few people really felt threatened by the disease. In this context, the behaviour within the household is arguably less likely to be direct consequence of the stress generated by the fear of the pandemic. Second, the quarantine decreed throughout Argentina established clear exceptions based upon labour categories as different as health care personnel, maintenance services workers, or retail supermarket employees (see Appendix for the full list of exempt activities). Therefore, for reasons exogenous to the prevalence of domestic violence, some families were placed in quarantine, while others were not. Our empirical strategy exploits in particular this second feature: the exogenous variability in the individual exposure to quarantine obligation to identify, using a difference-in-differences approach, the causal impact of quarantine on domestic violence. With additional information coming from the survey, the study also explores a series of potential mechanisms: (i) the increase in the time spent with family and partner; (ii) the negative shock to family income; (iii) the change in intra-household bargaining due to changes in the relative income of the family members; and (iv) the increase in alcohol consumption.