The recent spike in world food prices has intensified the debate regarding the impact of food prices on poverty. In this paper we aim to assess households’ vulnerability to food-price increases in four countries in Sub-Saharan Africa. Using household data from the World Bank’s Living Standard Measurement Surveys in Ghana (2005–2006), Kagera region, Tanzania (2004), Malawi (2004–2005) and South Africa (1993) we analyze food production and consumption patterns in rural and urban populations. In contrast to earlier studies we look at all food items and not just one or a few staple foods, giving a better understanding of vulnerability to general food price changes. We use two established indicators of sensitivity to food price changes—one measuring share of income spent on food, the other measuring elasticity of real income to food price changes. We find that the shares of the populations spending more than half of their income on food ranges from 62–81% in rural areas and from 26–67% in urban areas. Further we find that in all regions studied, most households (76–99%) in rural areas are net buyers of food and stand to lose in the short term from higher food prices. As expected, for urban households this is true to an even higher extent. Assessing the elasticity of real income to food price increases for the studied households we estimate that more than 25 million people in the four studied regions had their real incomes cut by more than a quarter in the 2007–2008 food price rise. Finally, we propose a new indicator, measuring the relative reduction in non-food spending needed to uphold food consumption at the same level after a food price change. We estimate that between 13–46% of the population in the four regions studied can not avoid reducing food intake and/or quality when food prices double, as they would have to decrease their non-food spending by 100% or more.