Abstract
Background
The current food, fuel, and financial crises have highlighted the importance of school feeding programs both as a social safety net for children living in poverty and food insecurity, and as part of national educational policies and plans.
Objective
To examine the costs of school feeding, in terms of both the absolute cost per child and the cost per child relative to overall education expenditure and gross domestic product (GDP) in low-, middle-, and high-income countries.
Methods
Data on the costs of school feeding in different countries were collected from multiple sources, including World Food Programme project data, reports from government ministries, and, where such searches failed, newspaper articles and other literature obtained from internet searches. Regression models were then used to analyze the relationships between school feeding costs, the per capita costs of primary education and GDP per capita.
Results
School feeding programs in low-income countries exhibit large variations in cost, with concomitant opportunities for cost containment. As countries get richer, however, school feeding costs become a much smaller proportion of the investment in education. The per capita costs of feeding relative to education decline nonlinearly with increasing GDP.
Conclusions
These analyses suggest that the main reason for this decline in the relative cost of school feeding versus primary education is a greatly increased investment per child in primary education as GDP rises, but a fairly flat investment in food. The analyses also show that there appears to be a transitional discontinuity at the interface between the lower- and middle-income countries, which tends to coincide with changes in the capacity of governments to take over the management and funding of programs. Further analysis is required to define these relationships, but an initial conclusion is that supporting countries to maintain an investment in school feeding through this transition may emerge as a key role for development partners.