Abstract
This study employs panel survey data combined with geo-coded information on household locations and civil conflict to examine the relationship between civil conflict and per capita food expenditure. Utilizing the panel structure and applying fixed effects models, the analysis shows that civil conflict and fatality rates significantly reduce per capita expenditure on home-produced food more than on purchased food (29% and 11%, respectively). Further disaggregation reveals that the negative effect of fatalities is more pronounced for per capita expenditure on both nutrient-dense and less nutrient-dense purchased foods, compared to their home-produced counterparts. Furthermore, our findings indicate that the diminishing effect of fatalities on expenditure for purchased and home-produced food remains robust across alternative estimation methods and varies across gender, age cohort, and landholding status. Importantly, maize price variability emerges as a key mechanism through which fatalities reduce households' consumption of both market-sourced and home-produced diets. The insights from this study provide policymakers and development practitioners with actionable strategies to strengthen household resilience in food consumption during post-crisis recovery. These findings also contribute to efforts toward achieving Sustainable Development Goal 2 (Zero Hunger) and Goal 16 (Peace, Justice, and Strong Institutions).