Violence against women: A cross-cultural analysis for Africa

Alberto Alesina, Benedetta Brioschi, Eliana La Ferrara

25 March 2016



Violence perpetrated by men against their female partners is one of the most common yet perplexing forms of violent behaviour. Besides being a fundamental violation of women’s human rights, domestic violence is becoming a significant public health problem, spawning high economic and social costs. Ripple effects throughout society can be enormous – victims of violence may suffer physical and psychological distress, and they may experience a decline in labour productivity and a resulting loss of wages, with limited ability to care for their children.

Domestic violence is not an isolated phenomenon. The World Health Organisation (2013) estimates that more than one third of women in the world have been victims of either physical or sexual violence, with low-income countries disproportionately affected.

The situation is especially worrisome in sub-Saharan Africa. Using the most recent data from the Demographic and Health Survey (DHS) for those African countries for which data on domestic violence are available, we estimate that 29% of women have experienced either sexual or physical violence since the age of 15. In addition, 46% of women in our sample justify wife beating, while the corresponding figure for men is 34%. What makes violence so widespread in Africa?

Economic and cultural factors that shape domestic violence in Africa

In a recent paper (Alesina et al. 2015), we investigate how economic and cultural factors influence current spousal violence in 18 sub-Saharan African countries.

  • We find evidence that the economic value of women affects men's violence against them. When ancient socioeconomic arrangements made women economically valuable, social norms developed in ways that viewed women as productive, more equal to men, and these gender roles bring about less intra-family violence today.

For example, having had bride price in the past is associated with a significant decrease in the probability and intensity of spousal violence.1 This suggests that if men traditionally had to pay for marrying their wives, they attributed a greater value to them and cared more about them. Interestingly, in line with this argument, the effect of lower actual violence seems to be driven by lower acceptance of wife beating on the part of the man.

In addition, having a pre-colonial environment that was more suitable for cultivating plough-positive crops is associated with a higher female acceptability of partner violence. This is consistent with Ester Boserup’s hypothesis, tested by Alesina et al. (2013), that differences in gender roles are influenced by historical agricultural practices. In particular, descendants of societies that practiced plough agriculture prior to industrialisation are characterised by more unequal gender norms today. This is due to the fact that the use of a plough required more physical strength, so in societies based on ploughs, women started to be relegated to the domestic sphere. Conversely, in a hoe culture, women were more likely to work outside the home and assume more important social roles, a feature that was then transmitted to their female descendants.

However, additional and subtle factors come into play. An economically more independent woman may also have more bargaining power within the marriage, which may lead to a negative reaction of men and ultimately to an increase – as opposed to a decrease – in violence. Indeed, when considering contemporaneous correlates of intimate partner violence, we find that if women currently work, spousal violence against them is higher. Interestingly, Bertrand et al. (2014) show that, even in advanced societies, intra-marital difficulties jump up discretely when the wife earns more than the husband. That is, when the wife ‘surpasses’ the man in terms of earning ability, there is a discrete negative psychological reaction of the man against the woman, holding everything else constant.

In addition, we find that being from an ethnicity that was traditionally endogamous (i.e. where members marry within the same ethnic group) has a positive and significant impact on spousal violence episodes, even when societies evolve. This may reflect less ‘modern’ cultural values of ethnicities which practiced endogamy in the past, or the possibility that beating a wife from a different ethnic group may bring about retaliation across ethnicities.

We also find that where the stem family (a small extended family in which two generations cohabit, as one son stays at the parental homestead with his wife and children) was socially predominant in the past, both men and women tend to have less favourable views of violence. This result is in line with the hypothesis advanced and tested by Tur-Prats (2015) for Spain. She argues that in stem families, wives have more time to work on the farm (while the mother-in-law takes care of the children), hence they become more ‘valuable’. Also the presence of two women in the household may deter the men from violence.

Matrilineality may matter too. Lowes (2015) finds that intra-family disagreements, which may escalate to physical violence, are more prevalent in matrilineal ethnic groups in Congo. In matrilineal ethnic groups, the woman may be relatively more empowered since group membership and inheritance are traced through female members.

Overall, these results suggest that the acceptance or rejection of violence against women is deeply rooted in ancestral social norms of different ethnic groups. These social norms persist even when economic and social conditions evolve, and may affect many generations (long-term persistence of cultural values has been widely documented).2 Given the pervasiveness of domestic violence against women, policy action is called for. Our results that violence behaviours are deeply rooted in ancestral socio-economic conditions warn us against disappointments from quick solutions.  

Concluding remarks

Cultural norms evolve slowly over time. Punishments per se may not be enough if society does not contemporaneously act to change dysfunctional social norms. A privileged channel is education (to the extent that educators do not perpetuate old norms). Communicating positive norms through the media and through modern social networks seems another promising avenue. More research is needed on the viability of different strategies to change cultural norms towards gender-based violence.


Alesina, A, BBrioschi, and E La Ferrara (2016), “Violence Against Women: A Cross-cultural Analysis for Africa”, NBER Working Paper, 21901.

Alesina, A and P Giuliano (2015), “Culture and Institutions”, Journal of Economic Literature, forthcoming.

Alesina, A, P Giuliano, and N Nunn (2013), “On the Origins of Gender Roles: Women and the Plough”, Quarterly Journal of Economics, 128(2), 469-530.

Bertrand, M, E Kamenica, and J Pan (2015), “Gender Identity and Relative Income within Households”m Quarterly Journal of Economics, 130(2), 571-614.

Boserup, E (1970), Woman’s Role in Economic Development. London, UK: George Allen and Unwin Ltd.

Lowes, S (2015), “Gender and Cultural Norms: Evidence form the Matrilineal Belt”, Working Paper, Harvard University.

Tur-Prats, A (2015), “Family Types and Intimate-Partner Violence: A Historical Perspective”, Working Paper 835, Barcelona Graduate School of Economics.

World Health Organisation (2013), "Global and regional estimates of domestic violence against women: prevalence and health effects of intimate partner violence and non-partner sexual violence", Geneva: World Health Organization, Department of Reproductive Health and Research.


[1] Bride price is a marriage practice which involves a payment in monetary terms or in kind to the bride’s family.

[2] For a recent survey see Alesina and Giuliano (2015)



Topics:  Development

Tags:  violence against women, Africa, domestic violence, cultural norms

Nathaniel Ropes Professor of Political Economy, Harvard University; and Research Fellow, CEPR

Analyst, The European House Ambrosetti

Professor of Economics, Bocconi University and CEPR Research Fellow