Association between food insecurity and intimate partner violence: the role of gendered asset policies

Abstract

Introduction: Intimate partner violence (IPV) affects an estimated 27% of women globally, with consequences spanning mental, physical and societal well-being. Previous research identifies individual and relational risk factors for IPV, but less is known about wider structural factors. This study examines the association between food insecurity and IPV using nationally-representative data from 156 countries, exploring gendered asset policies as a potential mediator.

Methods: We used nationally representative survey data on women's experience of IPV (1993-2019) and the Food and Agriculture Organization data on moderate to severe food insecurity. Multilevel mixed-effects generalised linear models estimated the association between standardised variables for food insecurity and IPV, accounting for country-level clustering and adjusting for gross domestic product (GDP). Mediation analysis estimated the role of gendered asset policies (Women, Business and the Law Index Score). Sensitivity analyses lagged food insecurity by at least 4 years among a subset of 59 countries.

Results: We examined data from 219 country-years representing 156 countries globally across the period of 1993-2019. After controlling for national GDP, models estimated an independent cross-sectional association between food insecurity and IPV (b=0.49; 95% CI 0.37 to 0.62). Gendered asset policies mediated 18.9% of the food insecurity-IPV relationship (indirect effect 0.099 (0.044-0.155); total effect 0.526 (0.422-0.631)). In lagged analysis, food insecurity was associated with higher IPV 2 years later (b=0.78; 95% CI 0.48 to 1.06), with a similar mediation effect (21.0%).

Conclusion: This study is among the first to harness global data to demonstrate country-level effects of food insecurity on IPV, which countries enacting more equitable asset policies were able to mitigate a substantial proportion of. Future research should prospectively pinpoint how supportive gender policies and asset ownership can amplify the benefits of food security for women's safety and longevity.

Keywords: Gender-Based Violence; Global Health; Health policy; Mental Health & Psychiatry; Public Health.

Figure 1. Path model presenting the hypothesised relationship between gross domestic product (GDP), foodinsecurity, gendered asset index, and intimate partner violence (IPV)Note: Independent variable (exposure), dependent variable (outcome), potential mediatingvariable, confounding variable; GDP; gross domestic product, IPV; intimate partner violence

Figure 2. Association between food insecurity and intimate partner violence (IPV) in unadjusted and adjusted models, including a model adjusting for gendered asset index as a variable. (a) Cross-sectional model: all 219 country-years. b) Lagged model: 59 country-years with multiple timepoints at least 4 years apart. Data on food insecurity and asset index from timepoint 1 and data on IPV from timepoint 2 (at least 4 years later). When included, GDP is baseline from at timepoint 1. Note: asset index; WBLI (Women, Business and the Law) gendered assets sub-index score; gross domestic product (GDP) based on purchasing power parity (PPP).

Figure 3. Path diagrams representing coefficients for direct and indirect effects of gendered asset index as a mediator between food insecurity and intimate partner violence (IPV). (a) Cross-sectional model: all 219 country-years. (b) Lagged model: 59 country-years with multiple timepoints at least 4 years apart. Data on food insecurity and asset index from timepoint 1, and data on IPV from timepoint 2 (at least 4 years later). Model controls for baseline GDP at timepoint 1. Note: path coefficients in dark blue box. Light blue box represents direct and indirect paths. *Significant association at p<0.05 level, ***significant association at p<0.001 level.