Conflict, food inflation, and climate change astime-varying determinants of food availability inSomalia: Evidence from machine-learning and VARtechniques

A new study published in Cogent Food & Agriculture examines how conflict, food inflation, and climate change have shaped food availability in Somalia over the past two decades. Using monthly data from 2001–2021, the researchers combine machine-learning models with vector autoregression and time-varying Granger causality to assess how the relative importance of these pressures has evolved. Across the full period, civil conflict emerges as the strongest predictor of food availability, followed by food inflation, temperature, and precipitation.

The study reveals an important structural shift in Somalia’s food system. Conflict accounted for about 70.7% of predictive importance during 2001–2007, but its share declined to 11.95% by 2015–2021. Over the same period, food inflation rose from 14.61% to 45.8%, becoming the most influential predictor, while temperature increased from 8.9% to 31.69%. These findings suggest that while insecurity remains a fundamental constraint, Somalia’s food system has become increasingly exposed to macroeconomic instability and climatic stress.

The findings carry an important policy message: addressing food insecurity in Somalia requires moving beyond a predominantly conflict-centred response toward a more integrated food-system strategy. Peace and security remain essential, but greater attention is also needed to food-price stability, import and market resilience, climate adaptation, domestic agricultural productivity, and protection against increasingly severe temperature and drought shocks. The study therefore provides new evidence for designing policies that respond to Somalia’s evolving hierarchy of food-system risks, rather than relying on assumptions based on past drivers alone.