The determinants of domestic violence: What can we learn from immigrants in Europe? Libertad González (Universitat Pompeu Fabra and Barcelona GSE) Nuria Rodríguez-Planas (Queens College CUNY) December 2017
Motivation What is the effect of “gender equality” on the incidence of domestic violence? Perhaps, if society is moving towards more gender equality, domestic violence will fall (social norms). But “male backlash” theories suggest that, when women do better in economic terms, violence may actually increase. More political empowerment among women may also increase their voicing of domestic-violence experiences.
What we do We use rich survey data for 28 European countries to learn about the determinants of the incidence of domestic violence. We focus on the role of culture, in particular social norms about gender equality. We start by documenting that women report higher levels of domestic violence in more gender-equal countries. We then explore whether women are more likely to under-report victimization in more gender-unequal societies. Then, we try to separate the roles of institutions, economic conditions, and social norms, by analyzing the incidence of domestic violence within country, focusing on immigrant women.
What we find We find that under-reporting is more prevalent in more gender-unequal countries. Evidence of empowerment channel. Women who come from more gender-unequal countries (within the same country of residence) are more likely to become victims of domestic violence. No evidence of “male backlash” channel. We argue that social norms (or “culture”) are important determinants of the incidence of intimate partner violence.
Contribution First paper to analyze the determinants of domestic violence with contemporary, high-quality, cross-country micro data for developed countries. First to use the “epidemiological approach” to study the impact of culture on domestic violence.
Literature Alesina et al. (2016), cross-country, descriptive evidence for Africa. Tur (2015, 2017), effect of culture (traditional family structure), Spain. Andenberg et al. (2015), effect of unemployment, UK. Aizer (2010), effect of the gender wage gap, US. Summarize the results in each! (see lit. review doc in November 2017 folder)
The 2012 FRA survey on violence against women Based on interviews with 42,000 women across the 28 member states of the EU. Who were asked about their experiences of physical, sexual and psychological violence. Including incidents of intimate partner violence (“domestic violence”).
Main measures of domestic violence Victim of intimate partner violence (IPV) ever (since age 15). Victim of IPV ever (confidential questions). Victim of IPV during past 12 months. Overall and physical/sexual.
Physical violence Your current or previous partner has: Threatened to hurt you physically Pushed you or shoved you Slapped you Thrown a hard object at you Grabbed you or pulled your hair Beat you with a fist or a hard object, or kicked you Burned you Tried to suffocate you or strangle you Cut or stabbed you, or shot at you Beat your head against something
Sexual violence: Your current or previous partner has: Forced you into sexual intercourse by holding you down or hurting you in some way. Attempted to force you into sexual intercourse by holding you down or hurting you in some way Made you take part in any form of sexual activity when you did not want to or you were unable to refuse Consented to sexual activity because you were afraid of what your partner might do if you refuse
Confidential questionnaire My partner or an ex-partner has been physically violent against me. My partner or an ex-partner has been sexually violent against me.
Overall incidence of domestic violence
Main measure of gender equality The “Gender Gap Index” (World Economic Forum) Gender gaps in economy, education, health and politics. Range: 0-1.
The GGI across EU-28 countries
The GGI and the incidence of domestic violence since age 15 (physical+sexual) (b=0.785***)
The GGI and domestic violence (confidential question)
The GGI and domestic violence in previous 12 months
Cross-country, individual-level regression results Dep. var: Incidence of domestic violence Indep. var.: Gender Gap Index Controls: Age, education, immigrant, married, children, working. (std. errors clustered by country) N=42,002 The regressions survive controlling for GDPpc (at the country level).
Incidence vs. misreporting The data suggest higher incidence of domestic violence in more gender-equal countries. This could signal higher real incidence (consistent with “male backlash” theories), or higher reporting. Can we say something about this? How people answer the regular questions by the interviewer vs. the confidential q’s in the self-completion questionnaire. Under-reporting: Answer NO to interviewer, but YES to confidential Q.
The GGI and under-reporting (b =-0.23** including all 28 countries, -0.14*** if we drop the outlier)
The GGI and misreporting, individual-level regressions
What’s driving the observed cross-country correlations? Differences across countries in: Institutions (laws, etc). Economic conditions (overall and for men vs. women). Differential misreporting. Culture/social norms.
Within-country analysis We focus on the sample of immigrants (including second generation). With country of residence fixed-effects. Holding institutions and economic conditions constant. The GGI in the country of origin as a proxy for culture. But culture could also affect misreporting?
Immigrant sample Confidential information on country of birth of mother and father. Immigrants and second generation. N=3,609 From 41 different source countries.
The GGI and domestic violence in previous 12 months
The GGI and under-reporting, within-country regressions
The GGI and under-reporting, within-country regressions Within-country, no evidence that “culture” is associated with under-reporting.
Within-country regression results Dep. var: Incidence of domestic violence Indep. var.: Gender Gap Index in source country (std. errors clustered by source country) List controls. N=3,609
Conclusions We study the association between the Gender Gap Index in the source country and the incidence of domestic violence among immigrants in the EU-28. With country of residence fixed-effects and individual-level controls. Using rich survey data from 2012. We find that women who come from more gender-equal countries are less likely to report victimization. Suggesting that more gender-egalitarian norms may reduce the incidence of domestic violence.
Conclusions (ii) We also find evidence of under-reporting of domestic violence in more gender-unequal countries. Which may be attributable to institutions, more than culture. Suggesting it may not be that hard to change. The evidence also suggests that the “male backlash” effect is probably not first-order.
Thanks for your attention!