Abstract
Despite growing availability of sex-disaggregated data on food insecurity, systematic cross-country evidence on the structural determinants of the gender gap in food insecurity remains limited. This study examines the evolution and macroeconomic determinants of the gender gap in moderate-or-severe and severe food insecurity between 2015 and 2023, using sex-disaggregated Food Insecurity Experience Scale (FIES) data from FAOSTAT's newly launched Gender in Agrifood Systems domain; the descriptive sample includes 154 countries and the regression sample includes 125. Descriptive analysis shows that the global gender gap widened during the COVID-19 pandemic, peaking in 2021 for severe and in 2022 for moderate-or-severe food insecurity, before narrowing through 2023. Latin American and Southern African countries display the largest gaps, with Afghanistan, Libya, and Botswana showing the most pronounced female disadvantage. A two-way fixed-effects panel regression with Driscoll-Kraay standard errors shows that, once the gender gap in agrifood systems employment is included alongside GDP per capita and agricultural GDP share, all three covariates are jointly informative: the agricultural share of GDP is positive and significant in both food insecurity models, GDP per capita is negative and significant for severe food insecurity, and the employment gap is positive and significant for severe food insecurity. A Hausman test confirms the fixed-effects specification is appropriate, and diagnostics rule out concerning multicollinearity. These findings suggest that structural dependence on agriculture and women's concentration in agrifood employment, rather than income level alone, are associated with wider gender disparities in food access, pointing to labour-market structure as an important and previously underexplored channel linking gender inequality to food access.