A natural disaster and intimate partner violence: Evidence over time

In India, living in a state affected by the 2004 tsunami was associated with higher rates of IPV immediately after the disaster. 


200 million people are affected by natural disasters every year, which are increasingly prevalent due to ongoing changes in climate. Recent research on climate-altering events has highlighted the distinct ways that climate events affect women, as natural disasters exacerbate social vulnerabilities that make women more susceptible to harm. Intimate Partner Violence (IPV) is one component of these gendered impacts of climate change. While research results are mixed, previous studies have suggested that disasters can heighten socioeconomic disadvantage and disrupt family dynamics, leading to an increase in IPV.

This study examines the rates of IPV in four Indian states, Tamil Nadu, Kerala, Andhra Pradesh, and Karnataka, before and after the Indian Ocean tsunami of 2004. This tsunami had devastating impacts across Asia, killing over 15,000 people. These states were affected by the tsunami to different degrees, with Tamil Nadu and Kerala being severely affected, Andhra Pradesh being moderately affected, and Karnataka being not directly affected. This study uses a natural experiment to evaluate how disaster severity is associated with IPV rates.  


Being from a state that was more severely impacted by the tsunami and having higher social disadvantage were associated with higher rates of intimate partner violence.  

State of Residence

  • In the 2006 and 2015 surveys, respondents from Tamil Nadu and Andhra Pradesh reported the highest prevalence of physical violence in the last year.
  • In the 2006 survey, 26% of women in Andhra Pradesh (which was moderately affected by the tsunami) and 28% of women in Tamil Nadu (which was severely affected by the tsunami) experienced physical, sexual, or emotional violence in the previous year.
  • Approximately 14% of women in Karnataka (which was not directly affected by the tsunami) and Kerala (which was severely affected by the tsunami) reported violence.
  • In the 2006 survey, respondents in all three states affected by the tsunami were also at higher risk of IPV compared to those residing in Karnataka.
    • Respondents from Tamil Nadu were 98% more likely to experience IPV compared to respondents from Karnataka.
    • Respondents from Andhra Pradesh were 85% more likely to experience IPV compared to respondents from Karnataka.
    • Respondents from Kerala were 41% more likely to experience IPV compared to respondents from Karnataka.
  • In the 2015 survey, respondents from Tamil Nadu were twice as likely to report IPV than respondents from Karnataka. 

Social Disadvantage  

  • In the 2006 survey, women with more social disadvantages were more likely to experience IPV in all four states.  
  • Women with a higher than secondary level of education were 47% less likely to experience IPV.
  • Having a partner with secondary education reduced the odds of IPV by 15%.
  • Being a member of disadvantaged caste groups was associated with higher rates of IPV compared to women belonging to more advantaged castes.  
  • Being a member of a minority religion (such as Islam and Christianity) was associated with higher rates of IPV.
  • Having a partner who consumed alcohol was associated with reports of IPV twice the rate of respondents whose partner did not consume alcohol.
  • In all three surveys, increased household wealth was associated with lower rates of IPV.
  • In the 2015 survey, caste and religious identity had no impact on IPV rates in all four states, although socioeconomic factors continued to contribute to IPV rates.  

The Indian Ocean tsunami of 2004 may have worsened rates of IPV, as evidenced by higher rates of IPV in states most affected. These findings also stress the importance of policies that center social inequities to ensure all groups are protected from IPV.  


This study utilizes data from the National Family Health Surveys from the years 1999, 2006, and 2015 to examine IPV before and after the Tsunami of 2004. Women were included in the study based on whether or not they answered domestic violence questions on the surveys. A total of 24,882 responses were included in the study across the three surveys.  

These data were used to complete three cross-sectional multivariable regressions. The dependent variable in this study was intimate partner violence, measured as a 0 (did not experience domestic violence in the last year) or 1 (has experienced domestic violence in the last year). Though the different surveys had different questions on IPV, the study coded any report of domestic violence through these survey responses to be a 1. The independent variable of interest in this study is the state of residence for each respondent, serving as a proxy for their closeness to the tsunami. Socio-economic factors (like education and wealth), demographic factors, and behavioral factors (including alcohol use) were also considered.  

Related GAP Studies