The World Health Organization has identified Southeast Asia as the region of the world with the highest rates of domestic violence, with 37.7 percent of women experiencing spousal abuse. This troubling statistic deserves the attention of policymakers and nongovernmental organizations looking to reduce domestic abuse. But those designing interventions should not treat this region as a monolith.

Recent research has highlighted that domestic violence is the result of community- and individual-level factors. Although certain socioeconomic groups, such as the impoverished and the poorly educated, are generally more likely to be the victims of domestic violence, the factors that put individual women at risk of abuse vary across communities.

Policymakers aiming to reduce spousal violence must be conscious of local context when designing interventions. Otherwise, policymakers risk using valuable resources on ineffective projects that do not address the root causes of domestic violence. Recent fieldwork by the Urban Institute profiles how different the causes of domestic violence can be, even among similar socioeconomic groups.

Where two cities align and differ

Last year, Urban’s partner, LEAD Pakistan, surveyed 1,200 men and women living in slums in Delhi, Dhaka, Islamabad, and Lahore. We collected data on various indicators that previous studies have connected with domestic abuse, including economic status, time use, women’s autonomy, and attitudes regarding gender issues like wifebeating and women’s labor rights. We also asked women if their spouses had physically abused them.

We found that Islamabad and Delhi were similar in traditional indicators of socioeconomic status, such as education, income, and employment. These cities also had similar levels of domestic abuse. Twenty percent of women in Islamabad and 23 percent of women in Delhi reported physical abuse.

But our analysis revealed that the most important risk factors were distinct in each city. In Islamabad, women with more autonomy faced a greater risk for physical abuse. Variables related to autonomy, such as decisionmaking authority within the family, financial independence, and progressive gender attitudes, increased the likelihood that a woman experienced physical abuse.

In contrast, in Delhi, most autonomy variables were not associated with reports of abuse. The exception was that women with more progressive gender attitudes were less likely to report abuse—the opposite of Islamabad. Further, variables related to women’s domestic and labor burden were important risk factors. Women who spent more time working and caring for children and who did not have paid leave were more likely to report domestic abuse.

The stark contrasts between the risk factors for the Delhi and Islamabad slums make more sense in light of the differences in the average levels of female autonomy between the two cities. Across most indicators of women’s autonomy—including family decisionmaking authority, financial independence, and attitudes toward wifebeating—Delhi ranked significantly higher than Islamabad.

Policy interventions must meet local communities’ needs

In the more culturally conservative Islamabad, women who disrupt the established social order through expressions of autonomy are at a higher risk for abuse. In Delhi, where high levels of autonomy are more common, women under high levels of stress from domestic and labor burdens are more at risk.

These two cities highlight the risks inherent in attempting to reduce spousal violence without a deep understanding of local context. At first glance, Islamabad and Delhi looked remarkably similar in terms of socioeconomic status. But an intervention to reduce violence by changing opinions about women’s autonomy at the community level would be more likely to succeed in Islamabad than in Delhi. On the other hand, an intervention aimed to reduce women’s domestic burdens would likely reduce domestic abuse in Delhi but not in Islamabad.

To reduce domestic violence in Southeast Asia, we must invest in interventions that work—projects that are flexible, adaptive, and meet local communities’ needs.