To assess, from a systems perspective, how climate vulnerability and socio-economic and political differences at the municipal and state levels explain food insecurity in Mexico.

Using a cross-sectional design with official secondary data, we estimated three-level multinomial hierarchical linear models.

The study setting is Mexico’s states and municipalities in 2014.

Heads of households in a representative sample of the general population.

At the municipal level, vulnerability to climate disasters and a poverty index were significant predictors of food insecurity after adjusting for household-level variables. At the state level, gross domestic product and the number of nutrition programmes helped explain different levels of food insecurity but change in political party did not. Predictors varied in strength and significance according to the level of food insecurity.

Findings evidence that, beyond food assistance programmes and household characteristics, multiple variables operating at different levels – like climate vulnerability and poverty – contribute to explain the degree of food insecurity. Food security governance is a well-suited multisectoral approach to address the complex challenge of hunger and access to a nutritious diet.